f9...
{.9
1?}...

.. 1wW...;

A.
u‘ .l — . . P
fixflri. r;

.u. é... .
:r

If y)...
:.l. ..
J.

 

 

 

 

 

-i}
s... 5-..;
uni- ..
A g '4 -.

 

r

: m I '53. 1' SQ) .M1~ . W,
a? a; mg? am}

an

A = I
1.... renal. u... 5..
. “a "u y .
:flnuzn Elm—fir «.1 .uh ‘I

5111......

W:

This is to certify that the
dissertation entitbd

TREATMENT EXPERIENCES AND OUTCOMES OF
AFRICAN AMERICAN WOMEN ADDICTED TO CRACK

PhD.

COCAINE

presented by

Janet Okagbue-Reaves

has been accepted towards fulfillment
of the requirements for the

degree in Social Work

 

Major Professor’s Signature
? - 2 3 — o 5

Date

MSU is an Affirmative Action/Equal Opportunity Institution

 

 

' LIBRARY
ftiniversiiy

 

 

—— -.-.-.-o--c-i--

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:/ClRC/DateDue.indd-p.‘l

 

TREATMENT EXPERIENCES AND OUTCOMES OF AFRICAN-AMERICAN
WOMEN ADDICTED TO CRACK COCAINE

By

Janet Okagbue-Reaves

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
School of Social Work

2005

ABSTRACT

TREATMENT EXPERIENCES AND OUTCOMES OF AFRICAN-AMERICAN
WOMEN ADDICTED TO CRACK COCAINE

By

Janet Okagbue-Reaves

This study is an analysis of treatment experiences and outcomes of African-
American women in substance abuse treatment for crack cocaine, and to identify factors
that contribute to their successful recovery as defined by completion of treatment and
substance abstinence one-year post treatment.

Data from a sub sample of 768 female clients treated in outpatient and residential
drug-free public sector programs, which were collected as part of the federally funded
National Treatment Improvement Evaluation Study (NTIES) (Kiecolt and Nathan 1985),
were utilized in this study. Positive treatment outcomes were identified by reports of
substance abstinence during the yearlong follow up period. Treatment components
examined included: length of stay in treatment, type of treatment program, and supportive
services received while in treatment. Client level factors examined included age,
employment status or sources of income, living arrangements, the raising of children,
criminal activity, social supports, marital status, and level of education.

The goal of the research was to identify factors within the sub—sample of African-
American women and within the treatment system that are associated with positive
treatment outcomes. Using logistic regression analysis, the study found that there is a

relationship between type of treatment program, length of stay in treatment, services

received while in treatment, and the women’s ability to remain drug free up to one year
post treatment discharge.

For the women who received services while in treatment, along with the type of
treatment program in which they participated, there was a significant difference in their
ability to remain drug and alcohol free one year post treatment. However, a majority of
the women in the sub-sample received little or no services while in treatment and had to
be excluded from the final analysis. Further research in this area would be needed to

capture a greater number of post-treatment outcomes.

This Dissertation is dedicated to my children, Michelle and Sasha Okagbue
(You are my sunshine)

and

to my Parents, William and Patricia Barr
(Mom, thanks for making sure I was always dressed appropriately)

and
George and Carolyn Logan
(Dad, thanks for never letting the rent fall too far behind and

for making sure we always had enough to eat)

For their untiring love, understanding, and support
Thank you

iv

ACKNOWLEDGEMENTS

The completion of this phase of my professional development is representative of
the interests and efforts of many people to whom I wish to express my appreciation. My
most profound feelings of gratitude are extended to all members of my immediate family,
who have, over the years, assumed more than a reasonable amount of responsibility to
facilitate this personal goal of achievement.

The interest, encouragement, and expertise afforded by the members of my
Committee, Dr. Teresa Jones, first Committee Chair, Dr. Victor Whiteman, second
Committee chair, Dr. Margie Rodriguez LeSage, Social Work, Dr. Clifford Broman,
Sociology, and Dr. Joanne Keith, Family and Child Ecology, is especially acknowledged
and appreciated. I would also like to acknowledge Sarra B. Barme for her assistance and
guidance in editing this document.

I would like to thank the MSU Graduate Office Fellowship Program for financial
support, and give special thanks to Dr. Francis and the Council for Social Work
Education for the Minority Fellowship Program which provided additional educational

and intellectual experiences, training, and support throughout this process. Thank you.

TABLE OF CONTENTS

LIST OF TABLES ................................................................................................... viii
LIST OF FIGURES ................................................................................................... x
CHAPTER ONE
INTRODUCTION ................................................................................................... 1
Research Questions ............................................................................................. 6
Intersections of Race, Class, and Gender ............................................................. 6
Ethical and Financial Considerations ................................................................... 9
Treatment Programs ............................................................................................ 1 1
Changes in Treatment Programs .......................................................................... 11
Substance Dependence and the Disease Concept ................................................. 12
Theoretical Models ............................................................................................. 15
Hypotheses ......................................................................................................... 19
CHAPTER TWO
LITERATURE REVIEW ......................................................................................... 22
CHAPTER THREE
RESEARCH DESIGN AND METHODS ................................................................ 30
NTIES Participant Sample .................................................................................. 31
How Individual Participants were Chosen .................................................... 31
Demographics of NTIES Population ............................................................ 31
Study Protocol .................................................................................................... 32
Data Collection ................................................................................................... 33
Full Sample ................................................................................................ 33
Study Sub-Sample ...................................................................................... 33
Validity and Reliability of NTIES Data ............................................................... 34
The Dependent Variable ...................................................................................... 35
Treatment Outcome .................................................................................... 35
The Independent Variables .................................................................................. 38
Type of Treatment Program ................................................................................ 38
Length of Stay .................................................................................................... 39
Ancillary Services During Treatment .................................................................. 39
Analysis of Data ................................................................................................. 40
Human Subjects Research ................................................................................... 41
CHAPTER FOUR
STUDY RESULTS .................................................................................................. 42
Study Participants ............................................................................................... 44
Treatment Factors ............................................................................................... 47
Dependent Variable ............................................................................................ 50

Analysis of Independent Variables: Types of Treatment Programs,

Length of Stay in Treatment, Services Received While in Treatment .................. 53
The Relationship between Treatment Modality and Drug Use ..................... 53
Respondents Who Took Crack, Cocaine, Marijuana, or Alcohol ................. 56
The Effect of Length of Stay on Drug Use .................................................. 57
Analysis of Ancillary Services Received during Treatment ......................... 66

Predictive Value of All Independent Variables on the

Dependent Variable ............................................................................................ 72

Sub-Sample of African-American Women One Year

Post-Treatrnent Discharge .................................................................................. 75

CHAPTER FIVE
DISCUSSION ......................................................................................................... 79

Type of Treatment Program ................................................................................ 79

Assessing Program Impact .................................................................................. 79

Length of Stay in Treatment ................................................................................ 82

Receipt of Ancillary Services .............................................................................. 84

Assessment of All Independent Variables on the Dependent Variable ................. 85

Limitations .......................................................................................................... 86

Future Directions ................................................................................................ 88

APPENDICES ......................................................................................................... 90

Appendix A: NTIES Demographics .................................................................. 91

Appendix B: NTIES Code Book Questions ....................................................... 96

Appendix C: Type of Treatment Programs ........................................................ 104

Appendix D: Variables Used for Analysis ......................................................... 106

Appendix E: UCHRIS Letter of Approval ......................................................... 108

REFERENCES ........................................................................................................ 110

vii

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7:

Table 8:

Table 9.

Table 10.

Table 11.

Table 12.

Table 13.

Table 14.

Table 15.

LIST OF TABLES

Identification of Referral Source as a Percentage of the

Sub-Population of African-American Women .......................................... 44
Identification of Reasons for Entering Treatment as a Percentage of the
Sub-Population of 641 African-American Women .................................. 45
Types of Housing One Year Pre-Treatrnent and One Year

Post Treatment ........................................................................................ 46
Number of Children Being Raised ........................................................... 46
Prior Treatment Experiences of the Sub-Population of 641

African-American Women ....................................................................... 47
Time in Treatment for the Sub-Population of 641

African-American Women ....................................................................... 48
Demographics for the Sub—Population of 641

African-American Women ....................................................................... 48
Drugs Used One Year Post-Discharge ...................................................... 50
Percentage of Women Who Used Poly Drugs One Year

Post-Treatment Discharge ........................................................................ 52
Types of Treatment Programs .................................................................. 54
Omnibus Tests of Model Coefficients of Types of Treatment

Programs on Drug Use ............................................................................. 55
Classification Table from SPSS ............................................................... 55

Percentage of Dependent Variable and Non-Drug Use, Attributed to
Independent Variable, Type of Treatment Program .................................. 56

Logistic Regression — Type of Treatment Program Analysis of
Maximum Likelihood Estimates .............................................................. 57

Duration of Treatment Episode in Months “ ALLDRUG

Respondents Who Took Crack, Cocaine, Marijuana, or Alcohol Cross-
stabulation ............................................................................................... 61

viii

Table 16.
Table 17.

Table 18.

Table 19.

Table 20.
Table 21.
Table 22.
Table 23.

Table 24.

Table 25.

Table 26.

Table 27.

Table 28.

Table 29.

Table 30.

Table 31.

Omnibus Tests of Model Coefficients for Length of Stay ......................... 63
Classification Table ................................................................................. 63
Percentage of Dependent Variable, Non-Drug Use, Attributed to
Independent Variable, Length of Stay in Treatment .................................. 63
Logistic Regression — Length of Stay Analysis of Maximum

Likelihood Estimates — Variables in the Equation .................................... 64
Services Clients Indicated at Intake that were Important to Them ............ 67
Services Received During Treatment of 641 Cases .................................. 68
Frequencies of Services Received During Treatment ................................ 68
Omnibus Tests of Model Co-efficients ..................................................... 70
Percentage of Dependent Variable, Non Drug Use, Attributed to
Independent Variable, Services Received ................................................. 7O
Classification Table ................................................................................. 71

Logistic Regression — Services Received. Analysis of Maximum

Likelihood Estimates ............................................................................... 71
Significance of the Overall Logistic Regression Model.

Omnibus Tests of Model Co-efficents ...................................................... 73
Predictive Value of the Independent Variables on Drug Use .................... 73

Percentage of Dependent Variable, Non Drug Use, Attributed to
Independent Variable, Type of Treatment Program, Length of Stay,
And Services Received ............................................................................ 74

Logistic Regression - All Independent Variables. Analysis of
Maximum Likelihood Estimates: Backward Stepwise .............................. 74

Demographic Information of Sub-Sample of Women One Year
Post-Treatment Discharge ........................................................................ 76

ix

LIST OF FIGURES

Figure 1.1. Urie Bronfenbrenner’s Ecological Theory of

Human Development ........................................................................... 17
Figure 4.1. Treatment Protocol Status .................................................................... 58
Figure 4.2. Length of Stay ..................................................................................... 60

Figure 4.3. A Cross Tabulation of all Drug Respondents Who took Crack,
Cocaine, Marijuana, or Alcohol ........................................................... 65

Figure 4.4. Sub-Sample of African-American Women One Year Post-

Treatrnent Discharge in 12—step Recovery Program .............................. 77
Figure 4.5. Employment Status of Sub-Sample of Women Pre-Treatrnent .............. 78
Figure 4.6. Employment Status of Sub-Sample of Women Post-Treatment ............ 78

CHAPTER ONE

INTRODUCTION

Drug addiction and treatment have historically been thought of as problems
concerning men, despite historical evidence to the contrary. Before the Civil War, the
number of female drug addicts outnumbered male substance abusers, and in the mid
18005, 60 to 75% of opium-morphine addicts were women (Wetherington & Roman,
1998). The 19905 marked the beginning of an increased focus on women’s health,
including addictive disorders, as inequities in women’s health research began to come to
light as more women entered health research fields.

For too long, research had been conducted by men with a generalization of results
estimated for women. It has since been acknowledged that research needs to be gender
specific and, if necessary, include racial and cultural differences. Researchers and
practitioners who study and treat women drug abusers need to acknowledge that this
group of women is not a homogeneous group. Adolescents, pregnant women,
housewives, older women, women from different ethnic groups, and lesbians may all
experience significant variations in the factors that lead to drug abuse and addiction. For
example, the effects of racism and community violence may contribute to making
African-American women more vulnerable to depression and substance abuse. In like
manner, a disproportionately large percentage of lesbians appear to suffer from drug and
alcohol disorders that may be related to their having to deal with negative societal

responses to their sexual orientation (Plumb, 2000).

Although this particular study focused on African-American women who are drug
abusers, it is important to point out that drug addiction is not a phenomenon that is
restricted only to the African-American community; drug addiction crosses all
socioeconomic boundaries and impacts multiple areas of society.

Drug and alcohol dependence is an increasingly devastating problem both for the
user personally and socially for the rest of society. In 2002, an estimated 19.5 million
Americans, or 8.3% of the population aged 12 or older, were illicit drug users.

0 Marijuana is the most commonly used illicit drug, with 14.6 million marijuana
users in 2002.

o In 2002, an estimated 2.0 million persons (0.9 %) were current cocaine users,
567,000 of whom used crack. Hallucinogens were used by 1.2 million persons,
including 676,000 users of Ecstasy. There were an estimated 166,000 current
heroin users.

0 An estimated 6.2 million persons, or 2.6 % of the population aged 12 or older,
were users of psychotherapeutic drugs taken non-medically. An estimated 4.4
million used pain relievers, 1.8 million used tranquilizers, 1.2 million used
stimulants, and 0.4 million used sedatives.

o In 2002, approximately 1.9 million persons aged 12 or older had used Oxy-Contin
non-medically at least once in their lifetime (Studies, 2002).

The economic cost to society from alcohol and drug abuse was an estimated $276
billion in 1995 (Abuse, 2003). According to statistics gathered in 2002, federal funding
for drug treatment totaled $3,587,500,000, representing 19.1% of the $18,822,800,000

drug war budget. The fiscal year (FY) 2003 request for treatment funding was

83,81 1,7 00,000, which represents 19.9% of the $19,179,700,000 requested drug war
budget. Combined, federal prevention and treatment funding totaled $6,136,100,000 in
FY 2002 with a requested $6,285,100,000 for FY 2003. Law enforcement and
interdiction consumed the remaining two-thirds of federal drug war spending (Policy,
2002)

An estimated 2.8 million people (1.3 % of the population ages 12 and older)
received some kind of drug or alcohol treatment in the 12 months prior to being
interviewed in 1999. Of this group, 1.6 million (0.7 %) received treatment for illicit
drugs, and 2.3 million (1.0 %) received treatment for alcohol. Treatment decreased
welfare use by 10.7% and increased employment by 18.7% after one year, according to
the 1996 National Treatment Improvement Evaluation Study (NT IES) (Center for
Substance Abuse and Treatment, 1996). There have been several research projects
recently completed using the NTIES data, which addressed the issue of identifying
factors that impact successful completion of treatment and abstinence from drug use.

The 1998 NTIES data found that, with treatment drug selling decreased by 78%,
shoplifting declined by almost 82%, and assaults declined by 78%. Furthermore, there
was a 64% decrease in arrests for any crime, and the percentage of people who largely
supported themselves through illegal activity dropped by nearly half. This study found
that treatment was a fraction of the cost of incarceration (Bureau of Justice Statistics,
1997)

Treatment for drug dependence has been shown to be effective in reducing crime
and health costs, yet many people needing treatment are unable to receive it. According

to recent studies, there are significantly more people addicted to drugs, who are in need

of treatment. The war on drugs has not significantly altered illegal drug use in this
country. More than 10 million people who need treatment each year are not receiving it
(US Dept. of Health and Human Services, 2000). A study by researchers at Substance
Abuse Mental Health Services Administration has indicated that 48% of the need for
drug treatment, not including alcohol abuse, is unmet in the US. (Woodward, 1997).

An important question for researchers is knowing whether the phenomenon of
drug addiction is different for women than it is for men. Men are more likely than
women to have opportunities to use drugs, but women, given an opporttmity to use drugs
for the first time, are equally likely to do so, and then progress from initial use to
addiction (Wetherington & Roman, 1998). However, women and men appear to differ in
their vulnerability to some drugs. Although both are likely to become addicted to
cocaine, heroin, hallucinogens, tobacco, and inhalants, women are more likely than men
to become addicted to or dependent on sedatives and drugs designed to treat anxiety or
sleeplessness, and less likely than men to abuse alcohol and marijuana (W etherington &
Roman, 1998).

In regard to treatment, many approaches to substance abuse treatment are
currently based on a narrow conceptualization of addictive behavior developed from male
experiences of substance abuse and recovery. Their treatment models may not be
appropriate for women and may even make women’s experiences with addiction appear
deviant and untreatable (Young, 2000). There are also differences between men and
women who seek treatment for drug abuse. Women in treatment programs are less likely
than men to have graduated from high school, to be employed, and are more likely than

men to have other health problems. They may also have sought previous drug treatment,

attempted suicide, and suffered sexual abuse or other physical abuse (Wetherington &
Roman, 1998).

The number of women using drugs, along with patterns of law enforcement, and
the severity of punishment for drug crimes increased the number of incarcerated women
by almost 500 % between 1980 and 1994 (Campbell, 2000). The proportion of women
incarcerated for violent offenses simultaneously declined. Women are not becoming
more serious criminals; rather, most women in federal prisons are there on drug offenses.

Specifically, women are affected to a greater degree than men, both socially and
personally, by drug addiction and alcoholism. Treatment services are limited for women
with few programs offering gender specific treatment. This gap in the provision of
services has become an issue of critical importance to drug addicted women who may
have specific co-occuning needs related to child custody, homelessness, employment,
domestic abuse, mental health issues, and crime.

These policies have disproportionately affected African-American, and nearly
70% of incarcerated women are single parents responsible for young children prior to
incarceration (Campbell, 2000). This research study took into account these factors
focusing on African-American women who are crack cocaine users, because of the drug’s
prevalence within low-income urban populations, which are often disproportionately
Afiican-American. The study was based on data from the NTIES research study, which
received public funding to target treatment centers in large urban areas, resulting in a
large percentage of African-American women participating in the study. Although there
has been previous research conducted on women crack cocaine users, this study’s unique

perspective took into consideration the demographics of the women participants,

especially that of single mothers with their children. Other considerations were housing,
familial support or non-support, and foster care agencies.

The primary interest in this research was to conduct an analysis of treatment
experiences and outcomes of African-American women addicted to crack cocaine and to
identify if the type of treatment program, length of treatment stay, and services received
while in treatment, impacted drug use one year post-treatment discharge.

Research Questions

The goal of this research project was to identify factors that contributed to
positive treatment outcomes for Afiican-American women who are addicted to crack.
Specifically, this project answered the following research questions:

1. Do variations in treatment protocols (e.g., type of treatment received)
influence treatment outcomes for African-American women addicted to
crack?

2. Does the type of treatment program, length of stay, and services received
impact recovery efforts?

By identifying the specific treatment factors that enhance success for crack
addicted Afiican-American women, this study hopes that this information can be used in
the future for recommendin g different treatment approaches.

Intersections of Race, Class, and Gender
It has been generally acknowledged that people of color in low-income groups
often do not have the same access to health care as those who are white and /or wealthier.
Lack of health care facilities, a dearth of doctors in low-income urban and rural areas,

and lack of medical insurance are cited as examples of the disparities facing people of

color (Patel & Rushefsky, 1999). This is a particularly compelling issue for women and
specifically for African-American women.

T reatrnent options and available health services for women have been shaped by
the patriarchal society in which we live. Women’s health studies have begun to bring
recognition to the fact that women have a broad spectrum of health needs beyond those of
reproduction, and that the provision of health services, historically, has been male gender
biased (Collins, 2000).

A central issue within the gender perspective is the notion of the sex-based
division of labor (associated with power relations) in which the division not only occurs
between the public and private spheres, but also within these spheres. The subordinate
valuation placed on women's activities and the inherent sexist values in the social
construct; ultimately translates into a differential allocation of family and social resources
needed for the promotion, protection, and maintenance of a woman's health (Rios, 1994).

We also need to be cautious about oversimplifying the reasons for drug addiction
that drive policy decisions. For example, conservatives may view substance abuse as the
product of personal weaknesses. They may argue that the problem of substance abuse
can be alleviated if people would take sole responsibility for their own actions and just
not do drugs. Liberals may see the problem as the result of a racially stratified social
system that distributes its resources inequitably. They may argue that the problem can be
resolved with a more equitable distribution of resources, such as jobs, housing, and health
care. But advocating for urban communities, the poor, and those addicted to drugs in a
conservative climate is simply not politic. Consequently, federal and state anti-dmg

budgets continue to support law enforcement over treatment and prevention.

What both liberals and conservatives ignore is that the social environment is one
constructed in arenas far from the inner cities in which a disproportionate share of poor
women are trapped (Zerai, 2002). There are myths and misconceptions about the poor,
about women, and about Afiican-Americans that affect policy and practice. A long-held
and common belief is that people are locked into a culture of poverty, crime, and drugs
because of their own behavior, thus alleviating society of responsibility for their welfare.

There are a disproportionate number of Afiican-American women in the prison
system with most incarcerated for drug related offenses. Initially, these women often
come into contact with the criminal justice system through a pregnancy when their drug
use is detected:

0 The overwhelming majority of the women charged with prenatal

crimes (about 80 %) are women of color and virtually all are poor.

0 Because poor women are under greater government supervision -
through public hospitals, welfare agencies, and probation officers -
- their drug use is more likely to be detected and reported.

0 Drug testing of newborns - the government's main source of
information about prenatal drug use - is implemented almost
exclusively by public hospitals that serve poor communities of
color.

0 Most states provide no statutory guidelines for testing and many
hospitals have no formal screening protocols, relying instead on
the suspicions of health care professionals. This discretion allows

doctors and hospital staff to perform tests based on their

stereotyped assumptions about the identity of drug addicts
(Studies, 2003).

For meaningful change to take place in improving the efficacy of drug treatment
for poor African-American women, environmental and systemic factors that impact a
person’s ability to fulfill their role expectations, must be taken into consideration.
Specific research designed to study sub-populations of women will be the only way to
develop successful treatment and prevention strategies to effect change in the targeted
populations.

Ethical and Financial Considerations

How do we as a society balance the needs of clients and the limited resources
available to assist them? Cost is a significant factor in the availability of services offered
within a treatment program. Across treatment settings, substance abusers present with a
variety of medical and social needs, and require comprehensive services to address these
needs. The lack of some services may either compromise a person’s ability to attend or
participate in treatment, or may lead to early termination. Early termination of treatment
is not cost effective and leads to poor clinical outcomes.

The larger ethical issues beyond cost effectiveness of treatment programs include
social as well as legal ramifications. Women with children have a particularly troubling
dilemma, since asking for substance abuse help can put them at risk of losing custody of
their children. In many states, habitual or addictive use of alcohol or drugs is evidence of
child abuse and neglect. The law’s tmfortunate effect is that children of substance
abusers are subject to permanent removal from their families as a result of several things:

their parent’s continued drug use fear of seeking treatment, and/or inability to access

appropriate treatment in their efforts to recover from addiction. Because substance abuse
is more likely to be detected among participants in social service programs for economic
aid and childcare, many of the children who are placed in foster care come from
disadvantaged households, headed by single, unemployed mothers who are
predominately minority.

The demographics of the system show that race influences state interventions in
families. African-American children enter the foster care system in grossly
disproportionate numbers, a disparity that has increased over the last two decades. By
1998, African-American children, who were only about 15% of the population under age
18, made up 45% of the nation's foster care population (Roberts, 2002). Eighty percent
of child welfare professionals surveyed said substance abuse exacerbates most cases of
child abuse and neglect they face (Child Guide, 2003).

The use of crack cocaine has had a particularly devastating effect on the drug
using population, and African-American women in particular. According to the
Substance Abuse and Mental Health Services Administration (2002):

o Fourteen percent of all adult female admissions to substance abuse

treatment were for the primary use of crack cocaine. About 58 % of
the adult female admissions for crack were African-American, 32 %
were white and 5 % were Hispanic.

0 In 2000, almost half of the adult women admissions for crack

cocaine reported alcohol abuse and 29 % reported marijuana abuse.

10

o In 1999, pregnant women aged 15 to 44 were more likely to enter
treatment for cocaine abuse than non-pregnant women of the same
age group.
0 In 1999, alcohol or cocaine abuse accounted for almost two thirds of
the 366,000 Afiican-American treatment admissions. African-
American female admissions were more likely to involve treatment
for "hard" drugs (e. g., opiates and cocaine) than were African-
American male admissions.
Types of Treatment Programs
The primary treatment programs for substance abuse are hospitalization,
outpatient, intensive outpatient, and residential. In outpatient treatment, the patient
remains in the community and is seen on an individual basis by a therapist from one to
two times per week up to one to two times per month. The next level of treatment is
intensive, outpatient treatment, which can consist of group and individual therapy two to
four times per week. Then, residential treatment is the next step up in treatment. Stays
range from two weeks to 90 days or possibly even a year for specialty programs.
Residential treatment is considered the treatment of last resort and is only for those
clients considered incapable of stopping drug abuse without being removed from their
environment. Inpatient hospitalization for drug treatment is primarily for detoxification.
Changes in Treatment Programs
With the rising cost of health care and a conservative shift in the societal view of
substance addictions, treatment options have been reduced and inpatient hospitalizations

reduced to three to five days of detoxification. Residential drug treatment days have been

11

slashed and the amount of money received from HMO’s and other health care
organizations and insurance companies greatly reduced. As a society, a greater and
greater portion of the responsibility for curing the disease has been shifted onto the
patient, which has fostered the development of shared responsibility for the disease.
Recognition of the multiple factors that impact the development and continuance of
addictive behaviors support the idea that treatment needs to be individualized.

Research has shown that many women in treatment face serious problems in
addition to their addiction, such as unemployment, problems with the legal system, health
problems, homelessness, and high rates of physical and sexual abuse (Wetherington &
Roman, 1998). Many of these women are of childbearing age and have concerns about
their children and their parenting abilities (Wetherington & Roman, 1998).

There is no single treatment approach for everyone, but particular approaches
need to be matched for everyone. From a budgetary point of view, the least restrictive
environment and the least expensive alternatives need to be attempted first before
comprehensive programming is called for in response to the diversity of client
characteristics.

Substance Dependence and the Disease Concept

As defined by the DSM IV, characteristics of substance dependence in an
individual may include the person taking the substance in larger amounts or over a longer
period than was originally intended. Other common characteristics of drug users are:

0 They may express a persistent desire to cut down or regulate substance use.

Often there have been many unsuccessful efforts to decrease or discontinue

USC.

12

0 They may spend a great deal of time obtaining the substance, using the
substance, or recovering from its effects. In some instances of substance
dependence, nearly all of the person’s daily activities revolve around the
substance.

0 Important social, occupational, or recreational activities may be given up or
reduced because of substance use. The individual may withdraw fiom
family activities and hobbies in order to use the substance in private or to
spend more time with substance using fiiends.

0 Despite recognizing the contributing role of the substance to a psychological

or physical problem the person continues to use the substance.

A person is dependent when they exhibit a maladaptive pattern of substance use
leading to clinically significant impairment or distress, manifesting three or more of the
above criteria in the same lZ-month period.

There is an ongoing debate in society over whether alcoholism and drug addiction
are diseases. Which side of the debate you fall on goes along way in predicting which
type of substance abuse treatment you would support. . Webster’s Dictionary defines
addiction as a compulsive need for and use of a habit-forming substance (as heroin,
nicotine, or alcohol) characterized by tolerance and by well—defined physiological
symptoms upon withdrawal; broadly used: “persistent compulsive use of a substance
known by the user to be harmful” (Merriam-Webster, 2002). The dictionary defines
disease as “a condition of the living animal or plant body or of one of its parts that

impairs normal functioning” (Merriam-Webster, 2002).

13

The disease concept of addiction suggests that there is a physical process
occurring in the addicted individual, which is beyond their behavior. Addiction is a
complex brain disease characterized by compulsive, at times uncontrollable, craving,
seeking, and use that persist even in the face of extremely negative consequences.
Scientific research has led experts to conclude that addiction is a disease or a chronic
illness like diabetes or hypertension. Some of the ideas of Stanton Peele (1989), who
strongly opposed viewing substance dependence as a disease in the early 1990’s, are still
prevalent today. Peele believed that treating alcoholism and other addictions as diseases
was a way for the addict to avoid personal responsibility and for the health care provider
to treat the addict indefinitely. Relapse is so common that it is used to justify the disease
idea. Failures are not evidence of the futility of treatment; people fail or relapse because
they can never fully recover from the disease (Peele, 1989).

Whether one views substance abuse and dependence as a disease or not, the
devastating effects on our society cannot be ignored. The United States is the largest
consumer of illegal drugs in the world. In 2002, an estimated 19.5 million Americans,
or 8.3 % of the population aged 12 or older, were current illicit drug users, up 2% from
1999 when an estimated 14.8 million Americans were illicit drug users or 6.7% of the
population twelve years old and older (Studies, 2002) . These factors have fueled the
federal war on drugs, mandatory drug sentences, and a conservative shift in the public’s
view of drug addicts and alcoholics. The populations that are dependent on public
funding for subsistence, shelter, food, and medical care are the most vulnerable when

public attitudes influence the shift of funding for public programs. As a result of public

14

perceptions of waste and fraud in recent years, there have been major changes in funding
for drug and alcohol treatment.

Considering all the information there is about drug dependence and lookin g at all
of the demographic issues, it is important to know what drugs are being used and to what
extent by various sub populations. What are the environmental and personal factors that
support or hinder a person’s ability to recover fiom drug addiction? What types of
treatment program and auxiliary services provide the best outcomes for different
populations of drug abusers? The statistics show that drug use and addiction are serious
problems for our society and the possible solutions merit detailed and comprehensive
research on personal and societal factors that can support recovery.

Theoretical Models
From Meta Yheory to Mid Level Theory

Quantitative and qualitative research efforts are best understood fiom within
theoretical frameworks, which can shape the direction of a study and provide a rationale
for the hypotheses on which the research is based. There are several theories and
perspectives that this research drew upon: the Ecological Systems Theory, the Strengths-
Based Perspective, and the Wraparound Process. Further, a feminist fiamework, which
suggests directions for change in social and environmental factors that create or
contribute to dilemmas and problems experienced by women on the intrapersonal and
interpersonal level, is a significant factor in the formulation of a theoretical fiamework
for this study. Feminist theory recognizes that researchers need to gather data concerning
family, fiiends, neighborhood, and other potential social supports as important sources of

information about strengths, circumstances, and perceptions for women. Researchers

15

need to examine individual circumstances and the influence of a person’s membership in
a particular socioeconomic class, ethnic group, or racial category (Lehmann & Coady,
2001). Comparing information about populations with comparable strengths and
challenges, which are affected by the same type of sociopolitical circumstances, will
enable researchers to determine if micro or macro interventions are necessary.

There was not one theory that best fit this study, but the wraparound process, a
service model for working with high risk populations, was applicable. The wraparound
process focuses on children and families with complex needs. These are children and
families who have needs in multiple life areas, which, if not met, may have a dramatic
negative effect on the quality of life of the family. The wraparound process is derived
from older generalist theories of social work practice such as systems theory and
strengths perspective.

The ecological systems theory deals with important societal processes (Bailey,
1994). Talcott Parsons (1952) theorized that a system exists when regularities of
relationships can be discerned among a set of parts and processes. The parts of a system
may be systems themselves. Any system that is not all-inclusive or closed will have
regular external relationships with its environment. The environment of one system will
also be a system if it has orderly internal relationships among its parts (Buckley, 1967;
Handel, 1993). Urie Bronfenbrenner ( 197 9), as well as other researchers, worked to
develop an ecological systems model used to describe how ecologies of smaller, micro
systems are affected by, and interact with, larger macro systems (Bronfenbrenner, 1979)

(see Figure 1-1).

16

   
  
 

Macrosystem

 

Mososystems: the interactions
among microsystems

 

A

\_ _/ V
Chronosystem: Patterning of environmental
events and transitions over the life course; socio-
historical conditions

 

 

 

 

Figure 1.1. Urie Bronfenbrenner ’s Ecological Theory of Human Development

The ecological systems model includes four levels: the microsystem, the
mesosystem, the exosystem, and the macrosystem. The first level is the microsystem
where individuals interact with their immediate family, school or place of employment,
neighborhoods, church, and peer groups. The second level is the mesosystem which
includes relationships between single settings, (i.e., local community mental health
organizations, the department of social services, law enforcement, or school boards) and
the individual directly involved. The relationships between settings are important in that
they can support an individual’s growth and firnctioning.

The third level, the exosystem, includes environments that may affect the

individual but with which the individual does not personally interact, such as the city

government. The fourth system is the macrosystem and includes broad ideological and
institutional patterns of culture or subculture that influence the development of cultural
norms that dominate the development in interactions of all systems, such as racism,
sexism, ageism, and homophobia GBronfenbrenner, 1979). Systems theory is broad and
does not specifically address sub—populations of society or disciplines in social work.
Social work is an applied profession and needs concepts that inform its work at the
concrete level of practice knowledge about addictions, child welfare, or mental health.

Based on the ecological systems theory, professional workers have been
encouraged to view systems holistically, attending simultaneously to clients and their
families, and whatever other systems may be important in meeting their clients’ needs.
This theory has encouraged professionals to recognize, on an ongoing basis, that the
various ecological levels are always reciprocally influencing each other (Lehmann &
Coady, 2001).

Another theory applied in this study is that of a strengths-based perspective which
challenges practitioners to perceive the strengths of people who often find themselves in
untenable situations (i.e., chemical dependency), and to engage these people in
identifying their strengths and encouraging them to use their strengths as tools useful in
achieving successful recovery outcomes (Rapp, 1997). The strengths-based perspective
is useful, but this study addressed treatment for African-American women at the meso
and micro level within the treatment agencies providing services. Thus, the wraparound
process, as a service model for meeting the treatment and psychosocial needs of African-

American women addicted to crack cocaine, is the most appropriate.

18

The wraparound process is concerned with systems of care and the development
of individualized service plans for families and individuals with complex needs, based on
their strengths. Even though the target population for the wraparound process is children
and families with complex needs, the focus is on strengths rather than problems, deficits,
and diagnostic labels. Wraparound service providers identify a child’s and family’s
unmet needs as the intervention target (Lehmann & Coady, 2001).

For the purposes of this study, the premise under-girding the wraparound process
is the philosophy used to formulate the hypothesis. Focusing on the needs of Afiican-
American women in treatment, the wraparound process was applied to the services
received while in treatment for these women, some with children at risk of removal fi'om
the home. The ecological systems perspective, combined with the focus of the specific
needs of individuals, as in the wraparound process helped assess the appropriateness of
' the services provided during and after treatment in the subpopulation of Afiican-
American women studied.

Hypotheses

Several hypotheses were put forward in this study.
Hypothesis 1

There will be an association between needed ancillary services received while in
treatment and the women’s ability to remain abstinent one-year post treatment. Ancillary
services are defined as life skills training and counseling while in treatment and

employment, housing assistance, interpersonal skills, academic, and family counseling

programs.

19

This hypothesis is supported by Feminist theory and the wraparound process.
Wraparound is concerned with systems of care and the development of individualized
service plans. Feminist theory also shapes this hypothesis.

Hypothesis 2

There will be an association between type of treatment received and ability to
remain abstinent one-year post treatment discharge.

This type of treatment program was defined by the Center for Substance Abuse
Treatment (CSAT) as a single site offering a single level of care. The classification of
level of care is based on three parameters: (1) facility type (e.g., hospital, etc); (2)
intensity of care (e. g., 24-hour, etc); and (3) type of service (e.g., outpatient, etc).
Previous studies have demonstrated that residential care, or 24-hour facilities, generate
the longest sustainable abstinence from drugs in the general population. This study will
test that assumption for this sub population of African—American women.

The theories supporting this hypothesis are the wraparound process and the
feminist theory, which support the idea of looking at the individual circumstances of
women and matching the intensity of services to their needs.

Hypothesis 3

There will be an association between length of stay in treatment and ability to
remain abstinent one-year post treatment discharge.

Length of stay in treatment is defined as the time spent in a treatment program,
either outpatient or residential. This study also looked at whether treatment was
completed. This hypothesis is supported by the strengths-based perspective and the

ecological systems theory. The issue under-girding this hypothesis is that of the

20

responsiveness of the treatment community to the needs of the women in the study. Are
the women in treatment for an appropriate length of time, and will differences in length

of treatment translate to different outcomes?

21

CHAPTER 2

LITERATURE REVIEW

The field of substance abuse has often been criticized for the lack of attention to
issues concerning African-American women’s substance use and abuse (Young, 2000).
There are even fewer studies in the existing literature in this field of research on African-
American women addicted to crack cocaine. Gender specific treatment is an area of
research that is continuing to develop. Within this area, research needs to focus on the
treatment experiences and outcomes of specific sub-populations of women. Few results
are available from studies that specifically are designed to assess interventions developed
for female substance abusers or to examine gender differences in the efficacy of existing
non gender-specific treatment (Wetherington & Roman, 1998). Some of the studies that
have been done will be reviewed in this section with an assessment of their relevancy to
African-American women, treatment experiences, and outcomes.

A recent study was completed using secondary data analysis on services received
during treatment and client outcomes. The authors found that the resumption of
substance abuse following treatment was more dependent on clients’ initial severity of
drug or alcohol use than on the intensity of services received during treatment (F inkbiner,
2000). This study was limited in that the original research was not designed to be gender
specific. Through the post hoc construction of three study groups—alcohol only, alcohol
plus other drugs, and other drugs only—systematic, between-group differences were
examined for client characteristics, client motivations for seeking treatment, specific

treatment modalities and services accessed, and the outcomes of treatment (F inkbiner,

22

2000). The study noted demographic characteristics of the populations studied in the
sub-groups but did not compare results across gender or race.

Other studies have also looked at treatment experiences and outcomes, but similar
to the Finkbiner (2000) study, they did not study specific sub-samples of the population.
Orwin (2000) conducted research on specific variables of treatment. This study looked at
the interactions between client-level factors and treatment components. Significant
interactions were discovered between hours in group or individual counseling, the
percentage of needed social/family services received, and the percentage of needed
employment! educational services received. Although women were significantly more
likely to be abstinent in most modalities, the positive effects of hours per month in
counseling, substance, and alcohol medications, percentage of needed services received
and treatment plans appeared to benefit men differently. This suggests that providers may
need to offer more tailored approaches to these services so that women benefit as well
(Orwin, 2000). Orwin’s study did look at a sub-group of women, compared to men, in
the study. The author was able to make some gender specific correlations with treatment
experiences, but he did not make specific correlations based on race.

In 2000, Comfort and Kaltembach examined whether client characteristics at
admission predict retention, abstinence, and utilization of required services and
specialized services for pregnant women in outpatient and residential substance abuse
treatment. Retrospective data, with psychosocial history, were collected. The
psychosocial history was administered to 133 women admitted to outpatient treatment
and 50 women admitted to residential treatment. Factor analysis reduced predictors to

five factors with composite scores, and multiple regression procedures determined client

23

characteristics that predict treatment outcomes. Using composites of variables reflecting
all aspects of women’s personal and family lives, it included medical and psychiatric
needs, family and parenting issues, housing, victimization, and clients’ perceived needs
for treatment and assistance in all of theses areas. The results indicated a need for more
holistic approaches to substance abuse treatment and continued exploration of a broad
range of psychosocial assessments at intake in order to develop substance abuse treatment .
programs that effectively address multiple aspects of women’s lives (Comfort &
Kaltembach, 2000).

Recent studies on the general population of women in treatment have shown that
women in short—term residential treatment complete treatment at a higher rate than
women in outpatient treatment (Gerstein, 2000). The percentage of drug users declined
by more than 40% during the year after the treatment episode compared with the year
preceding it, and average drug expenditures declined by about two-thirds. Arrests and
illegal activities declined by comparable amounts. There was some, but not dramatic,
improvement in employment rates and welfare dependency. Most studies focused on
pre-treatment variables and treatment experiences (Gerstein, 2000). Other studies have
found that women-only programs are the best way to address the special needs of women.
Studies have found that women in women-only treatment were more than twice as likely
to complete treatment as women in mixed gender programs (Scott, 2000).

Gerstein (2000) did gender specific research on women. Gerstein’s research
utilized the NTIES data set but did not go further than identifying demographic
characteristics of the study population and the sub-sample of women. The majority of

women in the research sample were African-American, in their twenties or thirties,

24

unemployed, single, and primarily in treatment for crack or cocaine powder. Many of the
women (40%) were raising children, and protecting or regaining the custody of their
children was a major reason for entering treatment. The women were treated mostly in
outpatient non-methadone, long-term residential, or short-term residential facilities.
Two-thirds of the women were on a second or later admission to treatment (Gerstein,
2000).

Gerstein’s research generated gender-specific information, but the research
needed to go a step fiirther and look specifically at the sub-sample of African-American
women, drugs of choice, and demographic characteristics of the women, not assuming
that they are a homogenous group. Of particular interest are women of childbearing age.

Several studies have cited the need for comprehensive substance-abuse treatment
programs that address the needs of women and substance abusing mothers and their
children (Klee, Jackson et al., 2002; Studies, 2003; Wetherington & Roman, 1998; Zerai,
2002). In psychosocial correlates of drop outs in a substance abuse treatment program for
mothers, Colenso (2002) looked at the relationship between psychosocial variables and
length of stay for women. She found that, although demand for services exceeded the
supply of spaces, early attrition rates within programs remained high. In the three year
study of a residential treatment program for mothers with a history of substance abuse,
participants answered questions regarding the association between length of time in
treatment and a number of psychosocial variables: level of depression, level of social
support, history of sexual/physical abuse, post-traumatic stress disorder, parenting

history, relationship with staff and peers, and parenting in family of origin (Colenso,

25

2002). Her results revealed that retention was associated with peer support, parenting
history, and an emotional empathy characteristic.

At the University of Michigan, Young and Boyd examined the relationship
between sexual trauma, severity of substance use, and treatment success for African-
American women addicted to crack cocaine. They interviewed 208 women who had
participated in a previous study. Group comparisons between women who had a history
of sexual trauma and those who did not were made with student t tests and Pearson’s chi-
square tests. Their findings indicated that sexual trauma is related to differences in
addiction and treatment experiences. Women with a history of sexual trauma had greater
difficulty breaking their addictions and had more previous treatment experiences. Their
research also indicated that there was a positive correlation between the severity of the
sexual trauma and the severity of drug use (Young, 2000).

In 2002, Reif examined whether ancillary services improve treatment
participation and substance abuse outcomes (abstinence) following treatment. Ancillary
services address non—substance abuse problems, such as employment, housing, education,
or medical care. It was hypothesized that treatment participation and outcomes would
improve when ancillary services were provided during substance abuse treatment, as
these services would reduce the impact of problems that trigger substance use or that
interfere with the treatment process. The sample consisted of 988 clients interviewed in
the Alcohol and Drug Services Study (ADSS), approximately one year after discharge
from outpatient non-methadone treatment. Very few clients received specific ancillary

services. Multivariate analyses generally showed that clients who needed, but did not

26

receive, each ancillary service did not have higher rates of abstinence than clients who
received it.

While contrary to the hypothesis, this finding suggests that clients with unmet
needs have the same potential for success following treatment as other clients. Treatment
participation was increased for clients who received specific services, but this did not
relate directly to abstinence at follow-up. Clients with the need for ancillary services had
somewhat different predictors of abstinence than clients with no need for ancillary
services, indicating the importance of individualized assessment. An enhanced focus on
identifying specific non-treatrnent needs may help to better determine the underlying
factors relevant to successful outcome. Although ancillary services did not predict post-
treatrnent abstinence in this study, further work needs to be done with other outcomes
such as employment (Reif, 2002).

A grounded theory of volition in recovery from substance abuse was studied by
Speck at the University of Nebraska in 2002. Her goal was to understand competing
influences in abstaining from alcohol or drugs. The study tied literature on motivation
from the field of education and the recent motivational constructs from addiction
literature to understand how goals and intention are protected. The data were gathered
through two sets of interviews and a reflection group using a grounded theory research
methodology. The results showed that maintaining recovery from alcohol and drugs is a
volitional process that involves comprehensive intrapersonal and interpersonal
developments in cognitive, physical, spiritual, and affective realms (Speck, 2002). Her
research further supports the need for addressing psychosocial issues for certain sub-

populations of drug abusing women.

27

Walton (1993 ), a graduate student at Michigan State University, studied
environmental factors and their impact on relapse. She noted that social and
environmental factors have not been extensively studied. Studies fail to consider the
social ecologies of the neighborhoods in which the families are embedded and the
communities in which leisure activities take place. Work, family, and informal support
networks were correlated to lower incidences of relapse (Walton, 1993). Her research
focused on assessing the validity of a setting-based relapse risk measure. She
interviewed 85 participants three times over the course of six months and found support
for the use of the measurement tool (Walton, 1993). More importantly, her research
identified the need for the development of instruments to measure environmental and
auxiliary factors and their impact on drug addicted women.

It is always important to look at people holistically and not assume that one type
of treatment program fits all. Funding for treatment should be spent effectively. A recent
study addressing benefit costs analysis demonstrated that investment in specialty
residential treatment for pregnant and parenting substance abusing women appears to be
economically justified given the long term social and economic savings to society when
these women are able to continue to care for their children (French, 2002). Many
approaches to substance abuse treatment are currently based on a narrow
conceptualization of addictive behavior developed from male experiences of substance
abuse and recovery. These treatment models may not be appropriate for women and may

even make women’s experiences with addiction appear deviant and untreatable (Young,

2000)

28

The literature has shown that significant differences in outcomes can be achieved
when treatment programs are tailored to women, those with and without children, and
towards differential treatments used for specific drugs. Treatment programs for women
substance abusers need to start with a model of female addiction that validates the
experiences of women. The pru'pose of this study was to look at those factors that
specifically impact Afiican-American women addicted to crack cocaine, a sub-population
not reflected in the current literature as being researched separately from women in
general. Even fewer studies have been completed that are specifically designed to study

women and, even more specifically, African-American women.

29

CHAPTER 3

METHODOLOGY

This study was based on a secondary data analysis of data from the National
Treatment Evaluation Study (NTIES) conducted from 1992 through 1997. NTIES is one
of the largest studies of substance abuse treatment ever performed in the US. NTIES’
primary purpose was to evaluate multiple federal grant programs in the area of substance
abuse treatment. In general, the grants targeted improvement of substance abuse
treatment services to under-served and vulnerable populations, including inner city and
public housing residents. Treatment programs, methadone programs, short and long-term
residential programs, and programs conducted within the correctional system, were the
source of study participants.

From a universe of 698 treatment sites, 82 were selected on a purposive basis for
participation in NTIES. Of the 82 sites, 78 were included in the final study. The
response rate among clients was 85% for a total of 6,593 clients participating in the
intake questionnaire. Clients were first interviewed around the time of their intake into
treatment. They agreed to be re-interviewed when they left treatment, as well as 12
months after treatment exit. Follow-up assessments included urine analysis drug testing

for about 50% of respondents.

30

NTIES Participant Sample

How Individual Participants Were Chosen

Treatment site admissions staff provided the interviewer with basic demographic
and primary drug data for clients as they entered the treatment program. The interviewer
would then solicit the client for participation in NTIES. In soliciting client cooperation,
interviewers explained the purpose of the study and the potential risks and benefits to the
client from participation, including extensive detail on the study’s confidentiality
protection. Clients provided written consent at each interview. Written consent was also
secured from a parent or guardian of any un-emancipated client.
The Number of Clients Who Participated

There were 6,593 respondents between July, 1993 and November, 1994 who
completed the intake questionnaire. All intake questionnaires were administered within
three weeks of the enrolled client’s first treatment session, using Computer Assisted
Personal Interview (CAPI) technology, at the NTIES treatment site.
Demographics of N TIES Population

The universe for the study consisted of those public sector substance abuse
treatment programs that were affiliated with a grant from the Center for Substance Abuse
Treatment (CSAT). These grant programs targeted under-served and vulnerable
populations including minorities living in inner cities and public housing. Targeted sties
consisted of substance abuse treatment programs in the United States receiving funding
from CSAT under one of three demonstration grants: Target Cities, Critical Populations,

and Criminal Justice.

31

Study Protocol

NTIES measured the outcomes of treatment primarily through a method known as
a "before/after" or "pre/post" panel design. From an available 698 Service Delivery Units
(SDUs) or treatment sites, 82 SDUs were selected on a purposive basis for participation
in NTIES. Of the 82, clients from 78 SDUs (for an SDU response rate of 95 %) were
included in the study. Clients were interviewed three times: shortly after their first day of
treatment, when they left treatment, and then at 12 months after the end of treatment. The
response rate among clients was 85 %, with 6,593 clients participating in the Intake
Questionnaire (NRIQ), 5,274 participating in the Treatment Experience Questionnaire
(NT EQ), and 5,388 participating in the Post discharge Assessment Questionnaire
(NPAQI

Records abstraction was completed for 6,420 clients. The records of those
respondents participating in all three interviews are flagged by the variable "IN_4411" in
the NPAQ data file (Part 3). This is referred to as the Outcome Analysis Sample and
includes 4,411 records, or 67 % of those participating in the initial interview. Some cases
were excluded from the analysis sample for reasons other than nonparticipation in the
three interviews, such as when the treatment exit date was missing or undetermined,
length of the interval for the follow-up interview was inappropriate (less than 5 or more
than 16 months) or the client was incarcerated for most or all of the follow-up period. Of
the SDUs sampled for the NTIES outcome analysis, 44 % were Target Cities programs,
38 % were Critical Populations programs, and 23 % were Criminal Justice programs.
Nearly half of the sampled SDUs were non-methadone outpatient programs, and about

one-quarter were long-term residential programs.

32

Data Collection

Data were collected across several important outcome areas, including drug and
alcohol use, physical and mental health, criminal activity, social functioning, and
employment. The study also collected information on participant’s treatment experiences,
and services received including, medical services, housing assistance, child care, legal
assistance, academic training, job training, family counseling, and group or individual
counseling.
Full Sample

The overall NTIES study consisted of 4,753 males, and 1,840 females. African-
American males numbered 2,442, Hispanic males numbered 729, and non-Hispanic non-
Afiican-American males numbered 1,582. Afiican-American females numbered 1,113,
Hispanic females numbered 245, non-Hispanic non—African-American females numbered
482. There were 4,411 participants who completed all phases of the study.
Study Sub-Sample

Of the more than 1,840 women who participated in the study, 1,113 were Afiican-
American non-Hispanic. Out of the 7 68 women in treatment for crack, 641 were
Afiican-American. There were broad age differences in the sample: 65.4% of Afiican-
American female participants were between the ages of 25 and 35, 11.3% were 17 to 24,
21% were 36 to 45, and 2.5% were 45 or older. There were 873 African-American
females who completed treatment. There were 641 African-American females in
treatment for crack addiction, 500 completed treatment. Out of the 641, 222 were in

outpatient treatment, 147 were in short term residential, 357 were in long term residential,

33

and 34 were in a correctional setting. This study looked at the treatment experiences of
the sub population of African-American women within each treatment modality.
Validity and Reliability of NTIES Data

The original researchers used several methods to ensure reliability and validity of
the study as detailed in the original report (Balnaves and Caputi 2001). The validity of
information reported by clients admitted to NTIES treatment sites, is an important issue
because much of the data used in NTIES to evaluate the treatment sites is based on
interviews with clients. For a variety of reasons, clients may over- or under-estimate their
drug use, criminal behavior, arrests, convictions, treatment needs, services received, and
other reports of their experience. Prior research comparing the self-reported data of
clients in treatment for alcohol or other drugs (AOD) with data from urine test results,
police records, and other sources found that client reports of their behaviors were
generally reliable and valid.

Reliability, whether or not a particular measurement technique applied repeatedly
to the same object will yield the same result each time, was ensured through the use of
established measurement instruments. Validity is the degree to which the observed value
captures the true value. In the NTIES analysis, the primary concern was validity that
measured to some extent the clients’ reports and records, drug tests, or administrative
data accurately and validly reflected whether a service was received, whether a client was
arrested, and whether a person used drugs.

Urine samples were sought from every individual interviewed at follow-up in
each treatment type until approximately half the estimated follow-up sample had been

interviewed, at which point the urine drug screening was ended. They also compared self-

34

reports of arrests with official records of arrests, obtained by matching identifying
information from NTIES respondents (but not other information from NTIES) with
identifying information in the statewide arrest report data files in states where the
respondents were enrolled in the study, and abstracting from any matched arrest records
information about the dates and types of offenses for which arrests were recorded (Center
for Substance Abuse and Treatment, 1996).

The Dependent Variable
Treatment Outcome

Treatment outcome was the dependent variable. Treatment outcome is defined as
the measurable change in drug use behavior that can be linked to participation in the
designated treatment program. Treatment outcome is measured by completion or non
completion of the designated treatment program and drug use or non-use one year post
treatment completion using the self report variable (T231) and chart abstraction variable
(F 1 8A). This study looked at the different types of treatment programs and compared
outcomes within the targeted sub-population of the study.

Many factors can affect whether people are able to maintain their sobriety once
completing treatment, including the type of treatment program they completed and
internal and external factors impacting their lives. Research has been conducted
analyzing various aspects of treatment including type of treatment, length of stay,
combination of treatment modalities, and the impact of case management (Inciardi, Tims
et. al., 1993).

A study done by Nirenberg and Maisto (1987) on treatment of alcohol concluded

that there does not seem to be strong evidence for the effectiveness of any treatment type,

35

but staying in any kind of treatment may increase the likelihood of improvement over and
above the rate of spontaneous remission. Studies that showed treatment effects were
unable to demonstrate long-term effects (N irenberg & Maisto, 1987). Unfortunately these
studies were on smaller populations and they made no distinction between residential
treatment and out—patient treatment, even though other research has demonstrated higher
rates of success with residential treatment. Other research has also tied positive outcomes
and drug abstinence to length of treatment and type of treatment (Gerstein, 2000; Orwin,
2000)

There is general agreement within the field that treatment improves outcomes for
the majority of people who are able to obtain it (Boyd, 1998; Sargent, 1992;
Wetherington & Roman, 1998). Researchers have struggled with the issues of how to
measure the effectiveness of substance abuse treatment. According to Stimmel (1983),
treatment in methadone maintenance, therapeutic communities, and outpatient drug-free
programs appears to be effective in diminishing recurrent drug use and antisocial
behavior (Stimmel, 1983). In a review of studies addressing individual and group
therapeutic approaches to the treatment of alcoholism, Stimmel concluded that patient
characteristics might be the most important determinant of response to therapy.

The absence of sufficient randomized control trials to study different treatment
modalities is an ongoing issue for the field of substance abuse and dependence treatment.
One of the common criticisms of randomized control trials is the establishment of control
groups, which deny persons, obviously in need of therapy, of appropriate psychological
support. Randomization and stratification often necessitate a multi-center study, which

may result in considerable variations in treatment rendered and accuracy of data

36

recorded. Aberrant conditions in a few centers might unfairly influence the results and
lead to erroneous conclusions; however, if an adequate study design is present, atypical
data should be identifiable and managed appropriately. The other difficulty noted with
randomized control trials and even general program evaluation follow-up is that of
assessing long-term efficacy when the actual outcome might be influenced by multiple
factors. A well-designed study should be able to identify confounding variables and
sequentially address these problems through a multivariate analysis (Stimmel, 1983,
1984)

Retention in treatment has been shown to have a positive relationship with
outcome. However, retention is also a problem. It has been estimated that up to 50% of
all users drop out of substance abuse treatment within the first month (Dilonardo, 1998).
Measures of the dependent variable include treatment completion and post treatment
abstinence. Research staff was concerned about possible discrepancies between self-
reports of drug use and actual drug use consequently the original researchers followed up
self-reports with drug tests:

To measure post-treatment abstinence, NTIES staff used self-reports and
successfully collected urine samples from 2,577 clients interviewed at follow-up.

There were only 19 outright refusals to provide urine samples by those who
completed interviews and approximately 85 cases in which obstacles (such as
telephone interviews, or a lack of privacy) precluded sample collection. Nearly all
samples yielded usable results from the NIDA-certified toxicology laboratory, which
tested for the presence, at standard cut-off concentrations, of chemical trace markers

of cocaine, marijuana, opiates (heroin, morphine, etc), methadone, amphetamines,

37

and alcohol. Urinalyses were compared to self-report data on drug use in the
previous 30 days that the NTIES staff had determined to use in the outcome analysis.
For cocaine, the drug used most frequently by clients, drug tests were positive about
29 % of the time; 30-day self-reports were positive only 20 % of the time; but self-
reports that referenced the number of uses during the follow-up period (with a
minimum of five) were positive more than 33 % of the time (Center for Substance
Abuse and Treatment, 1996).
The Independent Variables

Factors that impact the dependent variable, treatment outcomes, include type of
treatment program, length of stay, and services offered. The demographic characteristics
of the study participants included raising children, employment or sources of income,
level of education, and number of arrests pre- and post-treatment. Treatment background
factors included type of treatment completed, length of time in treatment, accessibility to
life skills training and counseling for employment, housing assistance, interpersonal
skills, academic, and family counseling. Personal factors were correlated with treatment
factors to determine if a relationship existed between treatment and personal factors and
treatment outcomes.

Type of Treatment Program

Type of treatment program was defined by CSAT as a single site offering a single
level of care. The classification of level of care was based on three parameters: (1)
facility type (e.g., hospital, etc); (2) intensity of care (e.g., 24-hour, etc); and (3) type of
service (e.g., outpatient, etc). A treatment program could be a stand-alone treatment

provider or it could be one component of a multi-tiered treatment organization. The

38

treatment methods used for this proposed study are outpatient, intensive outpatient and
residential treatment represented by (R2) treatment modality. .
Length of Stay or Length of Treatment Experience

Research has shown that length of stay is positively correlated with abstinence.
The study by Orwin (2000) showed a statistically significant relationship between length
of treatment and positive outcomes for non-methadone outpatient and long-term
residential modalities (McCoy, 2003; Orwin, 2000). Both completion of treatment and
length of time in treatment are associated with employment (McCoy, 2003). Length of
stay was measured by the number of days study participants remained in treatment
represented in the variable (LOS) length of treatment episode.

Ancillary Services During Treatment

Orwin (2000) found significant positive relationships between hours spent in
counseling, needed family/social services, employment/educational services received,
and participation in treatment planning. There is general consensus in the field that
providing ancillary services contributes to positive outcomes during treatment. Ancillary
services were measured by first looking at what respondents indicated at intake - if the
service was very or somewhat important to them; receiving help with housing problems
(R190); counseling or training about raising children (R248); counseling about family
problems (R250); counseling about getting along with others outside the family (R251);
treatment for emotions, nerves, or mental health problems (R311); counseling for
employment problems (R411); help or counseling for financial problems (R420); or
treatment for medical problems, child care (R100M4) and transportation (RIOOMI). The

study assessed whether services were offered, needed, and received.

39

Analysis of Data

The NTIES data were accessed through public domains specifically designed for
the purpose of encouraging further research using the data from the study. The data was
analyzed using SPSS statistical analysis software. The variables studied were
dichotomous variables (Maxfield & Babbie, 2001; Weinbach & Grinnell, 1995);
therefore the use of logistic regression analysis was appropriate for determining a
statistically significant relationship between the variables (Bachman & Paternoster,

1 997).

There are times when the dependent variable is not a continuously measured
variable with interval or ratio level properties, for example, dichotomous variables which
have only one of two outcomes. Dichotomous dependent variables can be more
accurately measured using logit or probit regression models rather than ordinary least
squares regression (Bachman & Paternoster, 1997). The logit and probit regression
models, like ordinary least squares, estimates a coefficient that measures the effect of one
or more independent variables on a dichotomous dependent variable. The dependent
variable in a logistic distribution is the natural logarithm of the odds of the dependent
variable occurring. The logit model helps to estimate a regression like equation that
contains coefficients that reflect the effect of each independent variable on the dependent
variable.

Personal background factors included race, type of drug used, history of arrests,
length of time using, education, age, social supports, raising children, employment status
and marital status. T reatrnent background factors included type of treatment completed,

length of time in treatment, accessibility to training and counseling while in treatment for

40

employment, parenting, housing assistance, interpersonal skills, academic, and family
counseling.

Potential difficulties and limitations included understanding that relationships
between treatment factors and outcomes is complicated by a number of factors: (1)
different populations respond to different treatment components, (2) it is difficult to filter
out the effects of specific components without experimental designs that isolate those
components, and correspondingly, (3) it is difficult to control for client self selection into
treatment services. This study was not a random study and its applicability to the general
population had limitations. Yet the relationship between social factors and positive
outcomes for African-American families is something that the treatment, policy and
human services field must understand better.

Human Subjects Research

Approval from UCHRIS was required and obtained for this study for a public
access data base. This was a research project involving secondary data analysis and the
issue of human subjects’ research was addressed in the original research. In compliance
with the Public Health Service Act, information obtained in the course of this SAMHSA
sponsored study that identifies an individual or entity was treated as confidential in
accordance with any promises made or implied regarding the use and purposes of the data
collected.

Access to raw data and machine readable files and any other personal identifiers
and other identifying or identifiable data collected were restricted to the principle
investigator and doctoral dissertation faculty committee members for assistance in

completion of the statistical analysis and interpretation of the data.

41

CHAPTER 4

STUDY RESULTS

This study was prompted by changes in the field of social welfare, especially in
that of child welfare, where recent legislation has limited the amount of time a woman is
given to meet court requirements and be re-unified with her children. The question we
must ask as a society, is, how can we better meet the needs of drug addicted women,
many of whom have children in the foster care system? The primary interest in this
research was to conduct an analysis of treatment experiences and outcomes of Afiican
American women addicted to crack cocaine and to identify if the type of treatment
program, length of treatment stay, and services received while in treatment, impacted
drug use one year post-treatment discharge.

This chapter will focus on the study results. First, descriptive statistics, including
frequencies and percentages, have been used to describe the demographic and clinical
pre-treatment characteristics of the study’s sub-sample participants. This includes
education, housing, previous treatment experiences, and illegal activities. After
describing each participant’s general demographic characteristics, the research questions
will be addressed, including the effect of ancillary services, types of treatment programs,
and lengths of stay in treatment for abstinence from crack cocaine and other drugs.

Drug use outcomes were measured by a dichotomous variable indicating any use
of crack cocaine, powder cocaine, and marijuana, heroin and alcohol post-treatment
completion. Secondary and tertiary drug and alcohol use was also analyzed due to the

high incidence of poly -substance use within the drug abusing population. Being able to

42

predict whether an event (drug use) will occur or not occur, as well as identifying the
variables useful in making the prediction, is important information for social scientists,
practitioners, policy makers, and academics.

A range of techniques have been developed for analyzing data with categorical
dependent variables, including discriminate analysis, probit analysis, log-linear
regression, and logistic regression (Schuessler 1971; Babbie, Halley et a1. 2000). Because
the dependent variable was categorical and dichotomous, the researcher chose logistic
regression analysis as the method for analyzing the data. Logistic regression analysis
calculates the odds of an event occuning and is defined as the probability of the event
occurring over the probability of the event not occurring. This equation indicates that the
dependent variable is the natural logarithm of the odds that an event will occur. The
regression coefficient is interpreted as the change in the natural log of the odds of the
dependent variable associated with a one-unit change in the independent variable. By
using multiple independent variables, this study was able to explain better, understand,
and predict the dependent variable (Bachman, 1997 #49).

The null hypothesis is that [31 type of treatment program, [32 length of time in
treatment and B3 services received equal zero association with drug use. The research
hypothesis was that the coefficients are greater than zero. This study determined if the
independent variables were significantly related to the dependent variable and then an
assessment of the relative effect of each determined which variable was more important

in explaining drug use or if the variables together had a greater impact.

43

Study Participants
The women in the study were fiom several different types of program settings,
with the majority coming from community-based programs in urban areas.
Approximately 36 of them, or 6% of the sub-population, were in correctional settings,
125, or 20%, were in public housing, and 480, or 7 5%, were in target cities. They were
referred from a variety of sources (see Table 1) and for a variety of reasons (see Table 2).
Multiple answers applied to individuals.

Table 1
Identification of Referral Source as a Percentage of the Sub-Population of

641 African-American Women.

 

 

Number Percent
Drug Treatment Staff 11 1.7
Criminal Justice Staff 38 5.9
Health Agency 5 0.8
School Teacher 2 0.3
Other Public Service 11 1.7
Employer 1 0.2
Spouse, Partner or Family Member 283 44.1
Friend 49 7.6
Other 16 2.5
Self 484 75.5

 

Respondents could choose more than one response.

44

Table 2
Identification ofReasons for Entering Treatment as a Percentage of the Sub Population

of 641 African American Women.

 

 

Number Percent
Criminal Justice Pressure 18 2.8
Partner/Parent Pressure 52 8.1
Teacher 1 0.2
Job 15 2.3
Parenting Issues 130 20.3
Health 26 4.1
For other services 20 3.]
Personal 588 91 .7
Other 65 10.1

 

Education, Housing, and Familial Responsibilities
Of the 641 African-American females in treatment for crack, 52.7% did not
complete high school, 29.5% had a high school education, 13.3% completed one or two
years of technical school and 4.5% completed three or four years of college. Housing,

prior to entering treatment, was also of interest (see Table 3 on next page).

45

Table 3

Types of Housing One Year I’m-Treatment and One Year Post—Treatment

 

Pr e-Tre atrn ent Post-Treatment

Number Percent

 

Number Percent

Own Apartment/Home 428 66.8 350 54.6
Other’s Home 395 61.6 255 39.8
Hotel 116 18.1 42 6.6

Homeless Shelter 168 26.2 76 11.9
Other Shelter 36 5.6 14 2.2

Hospital 108 16.8 37 5.8
Jail 119 18.6 66 10.3
Group or Half-Way

House/Trmt Ctr 412 64.3 137 21.4
Other Place 13 2 7 1.1

 

Only 121, or 19%, of respondents lived with a spouse or partner and 166, or 26%,
of respondents reported that their spouse or partner was a drug user. Approximately
119,0r 18.6 %, of respondents were married and 378, or 59%, lived with minor children,
55, or 8.6%, lived with adult children. The number of children that they were raising
varied (see Table 4).

Table 4
Number of Children Being Raised

 

 

N Number Percent
0 380 59.3
1 82 12.8
2 81 12.6
3 55 8.6
4 or more 43 6.7
Total 641 100

 

46

Compared to the original NTIES study where only 28% of the population was
raising children, the current study found that 41% of the sub-population of African-
Arnerican women was raising children.

This study showed that 205, or 32.2%, of the sub-population of Afiican-American
women were in treatment to regain custody of their children, and 230, or 36%, continued
to maintain custody of their children. Out of the women in treatment for crack, 77, or
12.3%, were afraid of losing custody of their children. When asked if they had an
emergency support person in their life, 587, or 91.6%, of the women responded
affirmatively.

Treatment Factors

Only 414, or 65%, of the women stated that they had prior drug treatment
experiences and of the 65%, 171, or 27%, had one prior treatment experience. The
remaining 38% had numerous prior-treatment experiences (see Table 5), and varying
lengths of time in treatment (see Table 6). There were specific characteristics that stood
out for this sub-group of Afiican-American women, including suicidal thoughts, rates of
shoplifting, and prostitution (see Table 7).

Table 5

Prior Treatment Experiences of the Sub-Population of 64 I A frican-A merican Women

 

 

 

N Number Percent
1 171 26.7
2 95 14.8
3 66 10.3
4 34 5.3
5 13 2.0
6 27 4.2

47

 

 

 

 

 

 

7-10 or more 1.2
Total 64.6
Missing. No prior
treatment 35.4
Total 100
Table 6
Time in Treatment for the Sub-Population of 64 I Africcm-American Women
N Number Percent
Less than 1 wk 23 3.6
1-2 weeks 50 7.8
3-4 weeks 108 16.8
1-6 months 174 27.1
7-12 months 32 5.0
More than 12 months 27 4.2
Total 414 64.6
Missing 7. No prior
treatment 227 35.4
Total 641 100
Table 7

Demographics for the Sub-Population of 64 1 African-American Women At Intake

 

 

 

 

 

Demographics

Number Percent
Attended 12-step program 554 86.4
Has had suicidal thoughts 174 27.2
Attempted suicide 204 31.8
Times Arrested
1 to 3 times 261 40
4 to 10 times 99 15

48

11 or more times 49 7

 

Times Had Sex for Money or Drugs

 

 

 

1 time 9 1.4

2 — 5 times 90 14

6—20 times 112 17.5
21 to 100 times 101 15.8
100+ 76 11.9
None 46 7.2

Times Shoplifted

1 time 29 4.5

2 —- 5 times 86 13.4
6 - 20 times 48 7.5

21 to 100 times 30 4.7

100+ 18 2.8
None 201 31.4

 

When asked if they were ever supported by illegal activities, 33.4% of the women
responded affirmatively. The demographics show multiple areas of need for this group of
women. Many of them are poorly educated, are not in their own homes or apartments,
have spouses or partners who are drug users, have children, are concerned about custody
of their children, have engaged in illegal activities, including exchanging sex for money
or drugs, and 95% had drunk alcohol in the past 30 days with 65% of the women stating

that alcohol abuse treatment was some what very important, to them.

49

The next section of this study will attempt to identify the similarities and
differences between this sub-group of Afiican-American women, their treatment
experiences, and ability to remain abstinent one year post-treatment completion.

Dependent Variable

At the end of the research project, 552 of the 641 women addicted to crack
cocaine participated in all phases of the study and 463 of the 641 women completed
treatment. In the post-discharge assessment, 39% (n=2 1 7) of all respondents reported
using crack in the past year. Conversely, 61% of the sub-sample (n=335) did not use
crack.

The analysis of the data attempted to determine a relationship between factors that
may have impacted how some of the women were successful in maintaining their
sobriety one year post-treatment discharge and others were not. Poly-substance abuse
within this sub-population of Afiican-American women was also looked at.

One of the consequences of drug addiction is the use of multiple substances. It is not
unusual for addicts to have one drug of choice, but use multiple drugs when their drug of
choice is unavailable, or engage in habitual use of drug combinations. Sobriety is defined
as abstinence from all drugs and alcohol. As shown in Table 8, the sub-group of African-
American women used a number of different substances.

Table 8:
Drugs Used One Year Post-Discharge

 

 

Drug Number Percent
Alcohol 251 45.6
Inhalants 3 0.6
Marijuana 1 18 21 .4

50

Crack

Powder Cocaine
PCP
Hallucinogens
Heroin
Methamphetamine
Narcotics

Uppers

Downers

Other

217
48

5

18

39.3
8.7

0.4
0
2.5
0.5
0.7
0.5
0.9

3.3

 

The drugs were examined next, other than crack, which were used by the sub-
group of African-American women (see Table 9). The table shows that there were
certain additional drugs used by the women who used crack, within one year of discharge
fiom the treatment program (N=217), and women who did not use crack (N=335). Poly
substances included marijuana, powder cocaine, alcohol, heroin, and PCP. The number
of poly drug users was significantly higher among the women who used crack within the
year after discharge than those who did not use crack. Within the sub sample, 314
women used one or more of the listed drugs. Analysis of this group of women shows that
217 or 100% of the women who used crack within one year also used one or more of the

listed drugs. Conversely, only 97 or 29% of the women who did not use crack, used one

of the listed drugs.

51

Table 9:
Percentage of Women Who Used Poly Drugs One Year Post- Treatment Discharge

 

 

Drug N % of Women N % of Women Who
217 Who Used Crack 335 Did Not Use Crack
and also Used: but did Use:

Alamo} 166 76% 85 26%
Inhalants 3 1% 0 0%
Marijuana 87 40% 31 9%
Powder Cocaine 35 16% 13 4%
PCP 2 1% 0 0%
Hallucinogens 0 0% O 0%
Heroin 9 4% 5 2%
Methamphetamine 2 1 % 1 .3%
Other Narcotics 4 2% 0 0%
Uppers 2 1% 1 0%
Downers 5 2% 0 0%
Other 11 5% 7 2%

 

Using a chi square test for significance, the following drugs were significantly
associated with the use of crack cocaine within this sub-population of women; Alcohol
p= 0.00, marijuana p= 0.00, and powder cocaine p= 0.00 were significant at the .05 level

In looking at the entire sub sample of 641 women, 39% used crack 61% did not.

If all drugs used are included in the analysis, then 29% of the women (see Table 9) who

52

did not use crack, used alcohol, marijuana, or powder cocaine which raises the percentage
of drugs used to 57% of the 641 respondents. Significantly, 43% of respondents used no

drugs or alcohol.

Analysis of Independent Variables: Types of Treatment Program, Length of Stay in
Treatment, Services Received While in Treatment.
The Relationship between Treatment Modality and Drug Use

Hypothesis 1. There will be an association between type of treatment received and
ability to remain abstinent from any drug including alcohol one year post—treatment
discharge.

Type of treatment program was defined by the Center for Substance Abuse:
Treatment (CSAT) as a single site offering a single level of care. The classification of
level of care is based on three parameters: (1) facility type (e.g., hospital, etc); (2) '
intensity of care (e. g., 24-hour, etc); and (3) type of service (e.g., outpatient, etc).
Previous studies have demonstrated that residential care, or 24-hour facilities, generates
the longest sustainable abstinence from drugs in the general population.

The theories supporting this hypothesis are the wraparound process and feminist
theory, which support the idea of looking at the individual circumstances of the women
and matching the intensity of services to their needs.

The women in this sub-sample were in fan different types of treatment settings:
short-term hospital, short-term residential, long-term residential and ambulatory

outpatient (see Table 10).

53

Table 10
Types of Treatment Programs

 

Percentage of Women Attending

 

 

Program Number Percent
Short-Term Hospital 26 4%
Short-Term Residential 87 14%
Long-Term Residential 325 51%
Ambulatory Outpatient 203 32%

N=641

 

Using logistic regression analysis, the overall model for types of treatment
programs as listed in Table 10, are significant for all drugs used (see Table 11). As
indicated by the classification table, which is one way to assess how well the model fits
compared to predictions of the observed outcomes, the model accurately predicted that
119 women, who did not use drugs, were correctly classified, but 208 women, who did
not use drugs, were incorrectly classified, for an accuracy rate of 34% (see Table 12)
(Agresti, 2002). The predictive value of the model is significantly more accurate for
women who did use drugs but is only accurate overall 54% of the time.

Rather than using a goodness-of—fit statistic, it is necessary to look at the
proportion of cases that were classified correctly. For this, the classification table
printed out by SPSS, indicates those cases where the observed value of the dependent
variable was 1 have been predicted with a value 1, and so on. An advantage of the
classification table is that at least one is found in either the logistic regression or the

discriminant analysis, so it can be used to compare the two approaches. Statisticians

54

claim that logistic regression tends to classify a higher proportion of cases correctly
(Psychology, 1997).

Table 11

Omnibus Tests of Model Coefficients of Types of Treatment Programs on Drug Use

 

 

Chi-square df Sig.

Step 1 Step 1 1.690 4 .020
Block 1 1.690 4 .020

Model 1 l .690 4 .020

 

Table 12
C lassification T able from SPSS

 

 

Observed Predicted: No — Predicted: Yes Percent Correct
Drugs Not — Drugs Used
Used ’
No - Drugs Not used 1 19 208 34
Yes - Drugs Used 82 232 73
Overall Percentage 54.8

 

55

Respondents Who Took Crack, Cocaine, Marijuana, or Alcohol

Next, the magnitude of the association between the type of treatment program and
the use of any drug was assessed. (Any drug is defined as crack, powder cocaine,
marijuana, or alcohol). The Cox & Schnell R2 and the Nagelkerke R2 Statistic are
statistics that attempt to quantify the proportion of the explained variation in the logistic
regression model. The Cox and Snell Statistic, when used in logistic regression analysis,
cannot achieve a value of one. In 1991, Nagelkerke modified the Cox and Snell Statistic
so that the value of one could be achieved (Norusis, 1999). In the model summary (see
Table 13), 37% of non-drug use is attributed to some type of treatment program.

Using binary logistic regression analysis to test for the significance of the
association between the dependent variable drug use and the independent variable type of
treatment program, the analysis showed that there was an association between drug use
and type of treatment program with 1: short-term hospital, 2= short-term residential, 3=
long-term residential, 4= ambulatory outpatient.

Table 13
Percentage of Dependent Variable and Non-Drug Use, Attributed to Independent
Variable, Type of Treatment Program

 

 

Step -2 Log Likelihood Cox &Snell R Nagelkerke R
Square Square
1 724.375 0.28 0.37

Using logistic regression, the treatment programs which were most significant in
the overall model were examined (see Table 14) (Menard, 2002). Treatment programs
included short term hospitalization, short term residential, long term residential and out-

patient treatment.

56

Table 14
Logistic Regression — Type of Treatment Program Analysis of
Maximum Likelihood Estimates

 

 

df Sig.
R2 Type of Treatment 4 0.02
Program
Short-term R2(1) 1 0.02
Hospital
Short-Term R2(2) l 0.01
Resid
Long-Term R2(3) 1 0.04
Resid.
Out Patient R2(4) 1 0.79
Constant Variable(s) 1 0.01
entered on step 1:
R2= Type a of
Treatment Program
(df) Degrees of Freedom
(Sig) Significance

 

All of the variables were significant except ambulatory outpatient treatment.
Short-term hospitalization, short-term residential and long-term residential treatment is
significant at the .05 level.

The Effect of Length of Stay on Drug Use

Hypothesis 2. There will be an association between length of stay in treatment
and ability to remain abstinent from any drug including alcohol one year post-treatment
discharge. Length of stay in treatment is defined as the time spent in either an outpatient
or residential treatment program. This study looked at whether treatment was completed.
The hypothesis is supported from both a strengths and an ecological perspective. The

issue under-girding this hypothesis, was the response of the treatment community to the

57

needs of the women in the study. Were the women in treatment for an appropriate length
of time, and did differences in length of treatment translate into different outcomes?

One of the outcomes in the study was the effect of length of stay in treatment on
the ability of participants to stay abstinent from drugs one year post-treatment discharge.
Statistical analysis showed that length of treatment stay had accounted for 10% of the
variation in the dependent variable. Treatment completion was also of interest. Using
descriptive statistics, the data indicated that 52%, or 332, respondents did not complete

treatment, but 42%, or 271 respondents either remained in treatment n=80 or completed

 

  
   

treatment n=191 .
Count
20° 4" Respondents who took crack, a rimming:
cocaine, marijuana, or alcohol
Took 1 or more
150 + of listed drugs
100 __
so __
0 ~0-

 

 

 

TX COMPLETED V rx NOT

INCONSISTENT
COMPLETED DATA

Figure 4.1. Treatment Protocol Status

Treatment completion is significantly related to drug use but the correlation is
small with only 10% of abstinence being explained by completing treatment. In

comparison to length of stay, which is for the most part under the power of the client,

58

treatment completion is a more complex outcome, involving providers to a large extent,
along with the client. Treatment completion can be seen as involving a judgment by the
provider that the client has made satisfactory progress and can now move forward to
maintain sobriety either independently or with less intensive assistance. Treatment
completion also involves client cooperation with the process and procedures required by
the treatment agency, and failure to complete drug treatment may often be linked to non-
compliance and early discharge by the provider, rather than reflect a measured judgment
about client progress.

The length of stay or length of treatment episode varied widely within this sub-
population of women. Approximately 45% of the study sub-sample stayed in treatment
one month or less, 13% stayed one to two months, 8% stayed three months 5% four - five
months, 7% six months, 3% seven months and about 14% the remaining months up to 24

months (see Figure 4.2 on the next page).

59

% of 641 Respondents

 

100%—it

75% -

I

5096—» 45%

 

 

 

0%—

1mo 1—2 3mos. 4-5 6mos. 7mos. 8-24
or less mos. mos. mos.

Figure 4.2. Length of Stay in Treatment

As you can see fi'om Figure 2, length of stay varied, with the majority of women
staying in treatment less than three months. Table 15 (on next page) details the length of
time in treatment for women who did not use drugs one year post-treatment discharge and

those who did.

60

Table 15
Duration of Treatment Episode in Months *ALLDR UG Respondents Who Took Crack,

Cocaine, Marijuana, or Alcohol _. A C ross-tabulation

 

 

 

Drug Use

Length of Stay No Yes Total
0.25 Count 19 37 56
% 5.8 11.8 8.8
0.5 Count 21 34 55

% 6.5 10.8 8.6
0.75 Count 31 32 63
% 9.5 10.2 9.9

1 month Count 31 84 115
% 9.5 26.8 18.0

1.25 Count 14.0 12.0 26.0
% 4.3 3.8 4.1

1.5 Count 12.0 12.0 24.0

% 3.7 3.8 3.8

1.75 Count 6.0 9.0 15.0
% 1.8 2.9 2.3

2 months Count 12.0 6.0 18.0
% 3.7 1.9 2.8

2.25 Count 5.0 4.0 9.0
% 1.5 1.3 1.4
2.5 Count 4 6 10

% 1.2 1.9 1.6

3 months Count 17.0 14.0 31.0
% 5.2 4.5 4.9

4 months Count 18.0 16.0 34.0
% 5.5 5.1 5.3

5 months Count 20.0 9.0 29.0
% 6.2 2.9 4.5

6 months Count 30.0 15.0 45.0
% 9.2 4.8 7.0

7 months Count 14.0 7.0 21.0
% 4.3 2.2 3.3

8 months Count 5.0 3.0 8.0
% 1.5 1.0 1.3

9 months Count 4.0 1.0 5.0
% 1.2 0.3 0.8

10 months Cormt 6.0 4.0 10.0

61

% 1.8 1.3 1.6

 

11 months Count 9 1 10
% 2.8 0.3 1.6
12 months Count 16.0 4.0 20.0
% 4.9 1.3 3.1
13 months Count 5.0 2.0 7.0
% 1.5 0.6 1.1
14 months Count 4 4
% 1.2 0.6
15 months Count 8.0 8.0
% 2.5 1.3
17 months Count 2.0 2.0
% 0.6 0.3
18 months Count 1.0 1.0
% 0.3 0.2
19-24 mos Count 3 3
% 0.9 0.5
20 Cormt 1.0 1.0
% 0.3 0.2
23 Count 3.0 3.0
% 0.9 0.5
24 Count 1.0 1.0
% 0.3 0.2
42 Count 3 2 5
% 0.9 0.6 0.8
Total 325 314 639
100 100 100

 

Using logistic regression analysis to test the association between drug use and
length of stay in treatment, the overall model is significant at the .05 level (see Table 16).
The classification table indicates that this model as an accurate predictor is accurate 61%
of the time for women who did not use drugs and 68% accurate for women who did use
drugs (see Table 17). The Nagelkerke statistic tells us that length of stay as a predictor

variable can account for 18% of a person’s ability to stay clean (see Table 18).

62

Table 16
Omnibus Tests of Model C oejficients for Length ofStay

 

 

 

 

 

 

 

Chi-square df Sig.
Step 1 Step 79.405 29 .000
Block 79.405 29 .000
Model 79.405 29 .000
Table 17
C [ossification Table
Predicted
Respondents Who Took Crack, Cocaine, Marijuana or Alcohol
Observed No Drugs Yes — Drugs Percentage
Used Used Correct
Step 1 No Drugs Used 201 124 61.85
Yes Drugs Used 100 214 68.15
Overall 64.95
Percentage
a The cut value is
.500
Table 18

Percentage of Dependent Variable, Non Drug Use, Attributed to Independent
Variable, Length of Stay in Treatment

 

 

Model Summary
Step -2 Log Cox & Snell R Nagelkerke R
likelihood Square Square
1 660.410 .134 .181

 

Using logistic regression analysis it was determined that length of stay (LOS) in

general is significant. There were several individual periods of time which were

63

statistically significant at the .05 level; weeks seven, twelve thirteen and fourteen and
weeks eighteen and nineteen(see Table 19).

Table 19
Logistic Regression — Length of Stay Analysis of Maximum Likelihood Estimates -

Variables in the Equation

 

 

 

df Sig.

Step 1 LOS 29 .001
LOS( 1) 1 .641
LOS(2) 1 .093
LOS(3) 1 .348
LOS(4) 1 .090
LOS(S) 1 . 179
LOS(6) 1 .662
LOS(7) l .018
LOS(8) l .222
LOS(9) 1 .71 1
LOS(10) 1 .060
LOS(] 1) 1 .078
LOS(12) 1 .003
LOS(13) 1 .001
LOS(14) 1 .012
LOS(15) 1 .133
LOS(16) 1 .075
LOS(17) 1 .128
LOS(18) 1 .009
LOS(19) 1 .001
LOS(20) 1 .073
LOS(21) 1 .668
LOS(22) 1 .544
LOS(23) 1 .761
LOS(24) 1 .830

64

LOS(25) 1 .710
LOS(26) 1 .830
LOS(27) 1 .710
LOS(28) 1 .830
LOS(29) 1 .262
Constant 1 ..018

 

a Variable(s) entered on step 1: LOS. LOS: Length of Stay

Figure 4.3 provides a visualization of length of stay correlated with drug use or
non-use. The figure shows that the longer patients stayed in treatment the greater the

number of non-drug users compared to drug users.

% of 641 Respondents

 

   

   

 

 

   

100% _ Missing or did not take drugs
- Did take one or more of listed
drugs

30%—

60% —

40%—

2091. —

0% .J 14.4
1 — 2 3

1m mos. 5m. 9mos. 15 24
orless mos. mos. mos.

Figure 4.3: A Cross Tabulation of all Drug Respondents Who Took Crack, Cocaine,

Marijuana or Alcohol

65

Again, overall, this model was significant at the .05 level with several individual
periods of time showing significant results, and in looking at the visualization of the
regression analysis the incidence of drug use is reduced if the individual remains in
treatment.

Analysis of Ancillary Services Received during Treatment

Hypothesis 3. There will be an association between ancillary services received
while in treatment and the women’s ability to remain abstinent from any drug including
alcohol one year post-treatment.

Ancillary services are defined as life skills training and counseling while in
treatment and employment, housing assistance, interpersonal skills, academic, and family
counsel. This hypothesis is supported by feminist theory and the wraparound process.
Wraparound is concerned with systems of care and the development of individualized
service plans. Feminist theory, which suggests directions for change in social and
environmental factors that create or contribute to dilemmas and problems, experienced by
women on the intrapersonal and interpersonal level, also shapes this hypothesis.

The independent variables concerning services received during treatment in this
study were looked at in two different phases. The first phase was to assess whether the
women indicated that there were areas of need in their lives, or if certain treatment
services were important to them. At intake there was a 99% to 100% response rate to the
variable questions during interviews. Table 20 indicates the women’s responses to

services they felt they needed.

66

Table 20

Services Clients Indicated at Intake That Were Important to Them

 

 

Vafiab‘e E13331 $332? 11.33am

Housing Help

n=641 N=135 21% N=46 7% N=460 72%

Child Care Counseling

n=640 N=90 14% N=86 13% N=310 48%

Social Problem Counseling

n=641 N=253 40% N=154 24% N=234 37%

Mental Health Treatment

n=640 N=l92 30% N=127 20% N=321 50%

Employment Counseling

n=641 N=199 31% N=114 18% N=328 51%

Financial Counseling

n=641 N=175 27% N=121 19% N=354 54%

Medical Treatment

n=641 N=127 20% N=111 17% N=354 63%

 

When asked about childcare problems, out of a total of 97 responses, 18 women
or 2.8% stated that it was a problem and 79 women, or 12.3%, stated that it was not a
problem. At discharge, 11 women, or 1.7%, had received childcare assistance. Only
3.7% of women indicated that transportation was a problem.

Table 21 details services received during treatment. This data is based on
information reported in the record abstraction from women’s files which indicated

frequencies of services received. (See Table 22).

67

Table 21

Services Received During Treatment of 641 Cases

 

Number of Services

 

 

 

 

 

 

N %
Received Housing Help F15A22 1-4 76 l 1.8
Received Childcare Counseling F15A15 1-4 195 30.4
Received Family Counseling F15A8 1-4 182 28.4
Received Social Problem Counseling F15A14 1-4 273 42.6
Received Mental Health Treatment F15A3 1—4 94 14.7
Received Employment Counseling F15A9 1-4 45 7.0
Received Financial Counseling F15A17 1-4 73 1 1.4
Received Medical Counseling F15A2 1-4 283 44.1
Received Child Care F15A21 1-2 1 1 1.7
Received Transportation F15A19 1-2 99 15.4
Table 22
Frequencies of Services Received During Treatment
Variable One Time 2-3 4-10 11 or More
# % # % # % # %

Housing Assistance

n=76 68 1 1 - 0 8 1.2 - 0
Parenting Skills

n=195 36 5.6 37 5.8 40 6.2 82 12.8
Family Counseling

n=182 26 4.1 35 5.5 39 6.1 82 12.8
Interpersonal Skills Training

n=273 25 3.9 61 9.5 72 11.2 115 17.9
Mental Health Services

n=94 66 10.3 23 3.6 3 0.5 2 0.3
Employment Counseling

n=45 10 1.6 13 2 5 0.8 17 2.7
Financial Assistance

n=73 42 6.6 25 3.9 4 0.6 2 0.3

68

Medical Services

n=283 95 14.8 86 _ 13.4 69 10.8 33 5.1
Transportation
n=99 16 2.5 34 5.3 30 4.7 19 3

 

All of the variables were entered into a logistic regression model using the
backward LR method. Initial analysis could not be performed because there were no
cases that received all of the services. Then the services were entered — those services
that were offered to roughly, 200 or more respondents. There were 47 valid cases that
received the targeted services. Focusing on the 47 cases included in the statistical
analysis, 44% of the dependent variable, non drug use, was attributed to the independent
variables, childcare counseling, family counseling, social problem counseling, and help
for medical problems. Help for medical problems was the least significant variable, so it
was removed and the statistic run again. All of the other variables entered were
significant at the .05 level. The next calculation had 96 valid cases with 34% of the
change in the dependent variable being explained by the remaining independent variables
(See Table 23). Though in this final calculation the category, received child care
counseling, was not significant at the .05 level.

There were several fiequencies that stood out from these two Tables. The first
was the discrepancy between needed housing and employment assistance, and the low
rate of housing and employment services actually offered. The second was the low
incidence of mental health services given that 50% of the women stated that it was very
important and 20% stated that mental health services were somewhat important. This
represented 7 0% of the entire sub-sample. In addition, this sub-sample also indicated that

27 % had had suicidal droughts and 32% had attempted suicide. Post-discharge, 15% of

69

them indicated that they had had suicidal droughts. Reported arrests were down to 7.5%
one time 11: 48, 2-3 times 1.9%, 4-10 times .02% and 11 or more times 03%. At follow-
up, 30% of the women were employed.

Other variables that were indicated as being important to the women at admission,
were either not significant or had too few cases to be able to be analyzed with any degree
of certainty, according to the record abstraction. Less than 100 women out of the 641
received housing assistance, psychiatric services, employment counseling, assistance
getting benefits, daycare for children, or transportation.

Table 23 demonstrates that the overall model using services received is significant
at the .05 level. In examining the Nagelkerk statistic 34% of non drug use can be
attributed to the services a woman receives (See Table 24). The predictive value of the
model is 77% (See Table 25).

Table 23
Omnibus Tests of Model Co-emcients

 

 

Chi-Square df Sig.

Step 1 Step 27.177 3 .000
Block 27. 177 3 .000

Model 27. 17 7 3 .000

 

Table 24
Percentage of Dependent Variable, Non Drug Use, Attributed to Independent

Variable, Services Received

 

 

 

Model Summary
Step -2 Log likelihood Cox & Snell R Nagelkerk e R Square
Square
1 95.034 .247 .342

 

70

Table 25

Classification Table

 

 

 

Predicted
No Drugs Yes — Drugs Percentage
Used Used Correct
Step 1 No Drugs Used 60 4 94
Yes Drugs Used 18 14 44
Overall 77.1
Percentage

 

The only significant variables in the regression model are family counseling, and
interpersonal skills training. Child care training had been significant in all of the earlier
calculations and was left in the model. (See Table 26).

Table 26
Logistic regression — Services Received.

Analysis of Maximum Likelihood Estimates

 

 

df Sig.

Step 1 F15A8 1 .045
F15A14 1 .017

F15A15 1 .072

Constant l .001

 

a Variable(s) entered on step 1: F15A8 =
Received Family Counseling, Fl 5Al4 =
Received Interpersonal Skills Training
F15A15 = Received child care Training.

Using logistic regression analysis the statistics demonstrate that interpersonal
skills training (F15A14) has the largest effect on not using drugs (see Table 26) and that

the model is significant at the .05 level with 34% (See Table 24) of the women’s ability

71

to remain abstinent from crack and other drugs being explained by the counseling

services received. Child care training was not significant in the final model.

Predictive Value of All Independent Variables on the Dependent Variable

Next an analysis was conducted using the complete model, which included the
following: length of stay in treatment, type of treatment program, services received while
in treatment, and their effect on drug use one-year post-treatment discharge.

The strength of the predictive value of the independent variables on any drug use
was analyzed. It was discovered that looking only at crack use does not provide an
accurate picture of the person’s ability to stay clean and the model is significant at the .05
level. Based on the Nagelkerke statistic, the independent variables have a predictive
value of 89% for a woman’s ability not to use crack, but a woman may not be using crack
alone, and may be using other drugs. Therefore, a calculation was made of the log of
odds of the independent variables to predict the dependent variable using all the
independent variables and the dependent variable of all the drugs and alcohol.

The overall model is significant at the .05 level (see Table 27). The predictive
value of the model for non-drug users is higher than drug users (see Table 28). Through
interpretation of the Nagelkerke statistic 40% (see Table 29) of a person’s ability to not
use drugs can be attributed to the combination of independent variables and their impact
on the dependent variable. Table 30 shows that all of the variables together have a
significant effect on the dependent variable, although type of treatment program is no
longer significant when combined with length of stay and services received while in

treatment.

72

Table 27

Significance of the Overall Logistic Regression Model. Omnibus Tests ofModel

 

 

Coefiicients

Chi-Square df Sig.

Step 1 Step 33.029 5 .000
Block 33.029 5 .000

Model 33.029 5 .000

Step 2 Step -.546 1 .460
Block 32.483 4 .000

Model 32.483 4 .000

 

Step 1 variables entered = type of treatment program, length of stay, and services received

Step 2 variables entered= length of stay and services received

Table 28

Predictive Value of the Independent Variables on Drug Use

 

 

 

 

Predicted
No Drugs Yes - Drugs Predicted
Used Used Percentage
Step 1 No Drugs Used 55 9 86
Yes Drugs Used 12 2O 63
Overall 78
Percentage
Step 2 No Drugs Used 56 8 88
Yes Drugs Used 16 16 50
Overall 75
Percentage

 

73

Table 29
Percentage of Dependent Variable, Non Drug Use, Attributed to Independent
Variable, Type of Treatment Program, Length of Stay, and Services Received

 

 

 

Model Summary
Step -2 Log likelihood Cox & Snell R Nagelkerk e R Square
Square
1 89.182 .291 .404
89.728 .287 .399

 

Table 30
Logistic Regression — All Independent Variables.
Analysis of Maxim um Likelihood Estimates: Backward Stepwise

 

 

 

df Sig.

Step 1 R2 1 .467
LOS 1 .015

F15A8 l .046

F15A14 1 .008

F15A15 1 .026

Constant 1 .021

Step 2 LOS l .016
F15A8 1 .051

F15A14 1 .009

F15A15 1 .023

Constant 1 .001

 

a Variable(s) entered on step 1: R2, LOS, F 15A8, F15A14, F15A15.
R2 = Type of treatment program

LOS = Length of stay

F15A8 = Received Family Counseling

F15A14 = Received Interpersonal Skills Training

F15A15 = Received child care Training

74

There were only n=96 valid cases for the logistic regression analysis and 545
invalid cases on one or more of the independent variables. The analysis shows that 40%
of the dependent variable drug use or non-use for all drugs can be predicted by the
independent variables for the valid cases. The small sample size is of concern, and
compromises the generalization of the results to the larger population.

When variables are removed, two things happen: the number of valid cases
increases and the effect on the dependent variable decreases, which demonstrates that,
together, all of the independent variables selected for analysis have a significantly greater
impact on being able to predict a person’s ability to stay clean.

I ran the statistics for women who lived with their minor children during one year
and for women who were afraid of losing custody of their children due to their drug use.
The findings were not significant for either variable, with less than 10 cases being used in

the analysis.

Sub-Sample of Afiican-American Women One Year Post-Treatment Discharge
I will begin the following discussion with basic demographic information
describing the sub-sample of African-American women one year post-treatment
discharge. When interviewed, the women showed improvement in several areas of their
lives as summarized in Table 31, and Figure 4.4, which details involvement in 12-step
recovery programs. Only n=33 or 5% of the women saw additional treatment providers

since their discharge date from the participating center.

75

Table 31
Demographic Information of Sub-Sample of Women One Year

Post- Treatment Discharge

 

 

 

 

 

 

 

 

 

 

Pre-Admission Post-Discharge

Number Percent Number Percent
Demographics
Attended 12-Step Prg 554 86.4 428 66.8
Had Suicidal Thoughts 174 27.2 80 12.5
Attempted Suicide 204 31.8 25 3.9
Times Arrested
l to 3 261 40 60 9.5
4- 10 99 15 1 . 0.2
11 or more 49 7 2 0.3
Times Had Sex for
Money or Drugs
One 9 1.4 4 0.6
2 — 5 90 14 13 2
6—20 112 17.5 25 3.9
21 - 100 101 15.8 17 2.7
100+ 76 11.9 18 2.8
Times Shoplified
One 29 4.5 5 0.8
2 — 5 86 13.4 21 3.3
6 — 20 48 7.5
21 — 100 30 4.7 7 1.1
100+ 18 2.8 1 0.2

 

76

120

 

Percent

 

 

 

°J

 

YES NO

Figure 4.4. Sub-Sample of African-American Women One Year Post-Treatment
Discharge in I 2-Step Recovery Program

An important component of maintaining sobriety for the sub-sample of African-
American women in a post-treatment program was to attend 12-step meetings. In the
pre-treatment assessment, 60% of the respondents attended five or more meetings and in
the post-treatment assessment, 97% of the respondents attended five or more 12-step
meetings. Figures 5 and 6 represent changes in employment status of the sub-sample of
Afiican-American women pre- and post-treatment. Slightly more than 7% of respondents

were employed pre-treatment, which increased to 29% post-treatment.

77

 

 

 

Fremency
O

 

YES NO

Figure 4.5. Employment Status of Sub-Sample of Women Pre- Treatment

 

 

 

Frequency

 

Figure 4.6. Employment Status of Sub-Sample of Women Post- Treatment

 

CHAPTER 5

DISCUSSION

This chapter will review the results of the statistical analysis completed in Chapter
four. This analysis included the use of binary logistic analysis to determine if there was
an association between the dependent variable, drug use and the independent variables
type of treatment program, length of stay in treatment and services received while in
treatment, separately and the combined effect of all of the independent variables on the
dependent variable.

Type of Treatment Program

The type of treatment program completed by the study participants was
significant at the .05 level in all treatment modalities except outpatient treatment. The
positive relationships found between the three treatment modalities of short term
hospitalization and short term residential and long term residential are effectively linear
across treatment durations (observed typically during single treatment episodes in these
modalities). The results from this analysis suggest that unusually long retention in long-
term residential and out-patient non-methadone programs, perhaps even when summed
across multiple episodes, are steadily less productive of improvement.

Assessing Program Impact

For addiction intervention studies, three general outcome domains have been
suggested to provide a thorough evaluation of program impact: 1) reduction of drug and
alcohol use; 2) improved personal and social functioning; and 3) reduced threats to public

health and safety (McLellan et al., 1996). These domains can be disaggregated further to

79

include any substance use, criminal activity, health services utilization, family and social
status, psychological status, or labor productivity.

In a research article by Zhang (2003 ), the author found that multivariate results
showed that dichotomous cut-points can generate significant differences in primary drug
use improvement in methadone maintenance, overall improvement in short-term
residential, and both overall and primary improvement in long-term residential. Their
base regression models, unadjusted for covariates or control variables, showed positive
linear relationships between duration and improvements in long-term residential and
methadone maintenance, supporting the common supposition that treatment duration is
an important predictor of outcomes (Des Jarlais, Joseph & Dole,l981; Simpson,198l; De
Leon,l984; Tims & Ludford,1984; Hubbard et al.,1989; Price,1997). However, their base
model also suggested that clients who stay in long-term treatment for extremely long
periods, (generally longer than 18 months), might start to experience reduced amounts of
improvement.

One of the measures for treatment success is being able to quantify the economic
benefit to society from people abstaining from drug use, but quantifying improvements in
psychosocial realms can be difficult. For example, psychological status is assessed
typically using clinical instruments such as the Beck Depression Inventory or the
Symptom Checklist (SCL) 90-R Global Severity Index. These instruments present
numerical scales for indicating degree of mental illness, and do not translate easily into
meaningful economic terms. One could infer that improved psychological status post-
intervention implies an economic impact, but quantifying this economic impact in

monetary terms is difficult.

80

Economic studies of addiction interventions have generally selected four outcome
domains from those mentioned above that are appropriate for monetary valuation:
criminal activity, health services utilization, productivity (employment), and substance
use. Various outcome measures exist within these domains that have been used to
estimate economic benefits.

For example, criminal activity may include self-reported criminal acts, arrest and
incarceration information, and policing and adjudication efforts. Health services
utilization includes medical and psychiatric services and may also include substance
abuse treatment. Employment outcomes, including type of job, number of hours worked,
and earnings generally assess productivity. Finally, substance use includes outcomes such
as frequency of use and expenditures on illicit drugs and/or alcohol, and may be included
in the estimation of addiction intervention benefits.

The appropriateness of this domain, as a true measure of economic benefits,
however, is questionable. From a societal perspective, the consumption of drugs and
alcohol is not in itself a direct cost. Society is, instead, primarily concerned with reducing
the social costs from the consequences of substance abuse, such as increased use of
emergency health care services or increased criminal activity to support a drug habit.
Because these measures of economic benefit are typically estimated independently,
including drug and alcohol use as a separate component to economic benefit, may lead to
double counting. However, some studies have included reductions in money spent on
drugs and alcohol as an important economic component indicating that those savings may

be used to buy other, less harmful goods. Of course, from the societal perspective,

81

expenditures on drugs and alcohol are actually an income transfer from one individual in
society to another and therefore do not represent a net benefit or loss.

My analysis demonstrated that all of the different types of treatment programs
were significant at the .05 levels except ambulatory outpatient treatment. The
predictability of the model in accurately classifying whether someone used or did not use
drugs based on the type of treatment program they were in was not a strong model with
less than a 50% accuracy rate for women who did not use drugs. Consistent with the
weak predictability of type of treatment program on drug use outcomes as shown in the
model summary, the type of treatment program was only accountable for 2-3% of
variability in the dependent variable.

Even though some of the literature supports the hypothesis that type of treatment
program has a significant impact on treatment outcomes, my research findings indicate a
weak relationship when looking only at type of treatment program.

Length of Stay in Treatment

Large-scale observational studies in the United States, such as the Drug Abuse
Reporting Program (DARP, based on admissions from 1969 to 1972), the Treatment
Outcome Prospective Study (TOPS, 1979-81), and the Drug Abuse Treatment Outcomes
Study (DATOS, 1991-93) have suggested that drug treatment outcomes are related to
treatment duration. Positive relationships between treatment duration and post-treatment
outcomes have been argued for each major treatment modality (Simpson 1979, 1981; Des
Jarlais, Joseph & Dole, 1981; De Leon, 1984', Tims & Ludford,]984; Hubbard et al.,

1989; Price,1997).

82

Consistent with some, but not all, previous studies, I found an association
between treatment duration and drug use outcomes. My statistical analysis found that
length of stay is a significant factor in a person’s ability to stay clean. There were several
individual periods of time which were statistically significant at the .05 level; weeks
seven, twelve thirteen and fourteen and weeks eighteen and nineteen. The differing
lengths of time, though significant, suggest opportunities for further research for
clarification as to what other factors may have lead to the differences in lengths of time
which are significant.

The classification model reflected moderately accurate predictability of drug use
depending on length of stay in treatment, and 13-18% of variability in drug use could be
attributed to length of stay. There was an association between length of stay in treatment
and drug use one year post-treatment discharge.

A considerable number of previous studies of the effect of treatment duration on
outcomes have used dichotomized measures (Gottheil, McLellan & Druley, 1992;
Hubbard et al., 1997; Simpson, Joe & Brown, 1997; Simpson et al., 1999). This approach
is sub-optimal in some respects. As most of these investigators recognize, many clients
with durations below, but near the cut-point may respond well to treatment, but for the
purposes of this study I chose to look at total abstinence which is optimal from a clinical
perspective. However, I do recognize that arbitrarily selected cut-point may be
misleading with regard to optimal durations for the average as well as any specific
individual. Ifthe threshold shifts downwards or upwards, similar claims may still be
drawn that those who stayed relatively longer than the 'threshold' would have better

treatment outcomes than those who stayed less than the threshold time-point.

83

Receipt of Ancillary Services

Participants in this sub-population of African-American women received fewer
services than they thought they needed and was not well matched to participant needs.
This finding was consistent with findings from the larger study. High levels of unmet
need for services were observed. This may suggest that substance abuse treatment
programs are limited in their ability to address the comprehensive problems of persons
entering substance abuse treatment. Other researchers have also reported sizable gaps in
the provision of comprehensive services to participants in substance abuse treatment
(Hubbard, 1989). Some investigators have suggested that such gaps are larger now, than
in the 1980’s and reflects the public’s decreased interest and support of substance abuse
treatment services.

Luchansky (2004), looked at employment outcomes, which was of interest to this
study given the high rate of unemployment within this population of women. He noted
that studies of employment after chemical dependency treatment, particularly those
employing secondary data, are rare. His work attempted to fill that gap. They documented
the employment outcomes, over a 42 year follow-up period, of a group with severe labor
market disadvantages. After the intervention, clients who completed treatment, and those
that completed a vocational program, earned significantly more than clients who did not.
These results were obtained even after controlling for differences in average wages in the
2 years before treatment. Differences in average earnings were consistent over the
follow-up period. Also, clients receiving treatment only, and those receiving vocational

services, earned significantly more after receiving these services than they did before.

My analysis determined that, thus far, services received had the greatest impact on
drug use one year post-treatment discharge for this sub-group of Afiican-American
women. The classification of the women was accurate at a 77% rate and the model
summary reflected that 32-44% of variation in the dependent variable is attributable to
services received.

Assessment of All Independent Variables on the Dependent Variable

The final analysis of my research involved the creation of a statistical model,
which incorporated all of the independent variables and their impact on crack use and the
use of all drugs one year post-treatment discharge. When all of the independent
variables were entered into the equation the number of valid cases was significantly
reduced. The sample size of 641 cases was reduced to n=96. The literature indicates that
this is not a highly unusual phenomenon and that using data from longitudinal studies can
lead to inconsistencies in the reporting of data as time passes and people may have
limited or distorted recall.

The inconsistencies in the data concerning services provided were also discussed
in the original survey. The original researchers discussed the interviewing techniques in
relation to the consistency of the data being reported. When asking study participants
about services received while in treatment, interviewers did not clarify the question for
respondents unless respondents specifically asked for clarification. Researchers also
noted that, in the record abstraction, services received were not documented consistently
across treatment centers.

Based on the limited number of valid responses, the combination of all

independent variables accounted for 40% of predictability in the dependent variable, or a

85

person’s abstinence from all drugs. The model was significant at the .05 levels.

In spite of the low number of valid cases, this research and my findings are
significant. Just as significant and of interest is the lack of services provided to women in
treatment despite, as in this case, the high level of perceived need for various services.
This project demonstrates that when services are received there is a significant
relationship between those services and a women’s ability to stay clean. The lack of
services highlights an area of need for this population.

A major advantage of the use of secondary data analysis for this research project
is that it was far less expensive than collecting raw data and had significant implications
for the development and testing of theories in the social sciences. Secondary analysis
allows for greater interaction between theory and empirical data because the transition
from theory development to theory testing is more immediate (Hakim, I982; Agresti,
1996). Also, the use of secondary data analysis may lead to the development of new
approaches to methodology, such as the use of replication tests to substitute for, or
complement, significance tests, a useful feature of secondary research. Further, testing
the same hypothesis on two or more data sets of the same type can lead to greater
significance and reliability than the use of a single data set (Hakim, 1982). This study
was limited by the use of only one data set. 4

Limitations

This study was subject to several limitations. 1 assessed the effects of one
treatment episode only, while it is well known that treatment for drug and alcohol
addiction often involves more than one episode. It is quite probable that treated clients

would have received additional treatment in the follow-up period, or that they received

86

services from an additional source. If true, then those additional services might explain
the group differences I observed. Unfortunately I had no data about services received in
the follow-up period. However, the probability of obtaining additional services was
equally great for those receiving only small amounts of treatment. For this reason, it is
impossible to tell what influence these missing data might have had on the results.

Relying on an archival data source allowed for a long follow-up, and for
particularly good data on the outcome measure. However, I did not have data over time
on client characteristics. The ideal study of this sort would combine objective outcome
data with self-reported measures on important client characteristics taken at regular
temporal intervals. However, because of attrition, such a study would be very difficult to
conduct.

Finally, because of the quasi-experimental design, these findings should not be
taken as conclusive. Random assignment was not possible. The CSAT intent is to offer
treatment to all that are deemed clinically eligible, as opposed to creating conditions for
conducting rigorous, controlled experiments. True random assignment has the advantage
of precluding any selection effect, but as Holder et al., (1992) point out, even the
diagnosis of a substance abuse problem introduces selection, because in most cases
clients must present themselves for diagnosis. It is possible to randomize clients after
they elect to pursue treatment, but randomization before that time, when it would be most
appropriate from a scientific point of view, is nearly impossible. Also, even if random
assignment was more feasible, some clients who want treatment might have chosen to

avoid a random assignment situation, again introducing a selection effect.

87

Future Directions

With an increase in services, consistency in time spent in treatment, and greater
availability of long-term residential treatment, society can help improve treatment
outcomes for women addicted to crack. This study helped highlight the need for the
development of research instruments to better assess the life situation, social support
networks, skills, and areas of need for African-American women drug users.

The social and economic cost of drug addiction can be tempered with precise
targeted interventions by service providers based on clinical and social evaluations of
clients. The findings of this study, which has predicted greater abstinence from drug use
with services matched to clients needs, participation in long-term residential treatment,
and appropriate length of stay, is significant for a number of reasons even though the
results are based on a small sample size. The positive relationship between length of stay
and treatment outcome for users of illicit drugs has been consistently demonstrated
(Simpson, 1995: Gersteirn, 1990), and suggests that a way of improving client outcomes
is to provide services matched to client-perceived needs. Also, many in the substance
abuse treatment field have endorsed the provision of comprehensive services for
substance abuse treatment participants, which, until recently, has been lacking. This is
particularly so for clients treated in outpatient drug-fi’ee programs.

This study has wide implications for policy makers and treatment providers and
can assist in focusing resources where interventions can provide a greater return in
abstinence from drug use and increased productivity of members of our society addicted
to drugs. This study has even wider implications for providers of child welfare services

who are working towards family reunification for parents with substance abuse related

88

issues. With greater limitations on the amount of time allowed a parent to meet court
requirements for reunification with children in foster care, this research enables providers
to better predict outcomes when services, type of treatment program, and length of stay in
treatment are matched to client needs.

This study focused on African-American women and acknowledged that efficacy
of services is an issue that must be addressed to meet the needs of this population of drug
addicts, and to contain treatment costs. To effect meaningful change in improving the
efficacy of drug treatment for poor African-American women and all women of color,
environmental and systemic factors that impact a person’s ability to fulfill their role
expectations must be taken into consideration. Specific research designed to study sub—
populations of women will be the only way to develop effective treatment and prevention

strategies to effect change in the targeted populations.

89

APPENDICES

9O

APPENDIX A

NTIES Demographics

91

NTIES Demographics

NTIES is a congressionally mandated five-year study of the impact of drug and
alcohol treatment on thousands of clients in hundreds of treatment units that received
public support from the Substance Abuse and Mental Health Services Administration
(SAMHSA), Center for Substance Abuse Treatment (CSAT). The NTIES project
collected longitudinal data on a purposive sample of clients in treatment programs
receiving CSAT demonstration grant funding. Client-level data were obtained at
treatment intake, at treatment exit, and 12 months after treatment exit. Service delivery
unit (SDU) administrative and clinician (SDU staff) data were obtained at two time
points, one year apart.

Data were collected across several important outcome areas, including drug and
alcohol use, physical and mental health, criminal activity, social functioning, and
employment. For a random sample of approximately half of those interviewed, urine
specimens were collected at follow-up to corroborate clients' self-reports of substance
abuse. Substances covered in the study included alcohol, analgesics, anti-anxiety
medications, anticonvulsants, antidepressants, anti-manics, barbiturates, cocaine (powder
and crack), depressants, hallucinogens/psychedelics, heroin and other opiates, illegal
methadone, inhalants, marijuana/hashish, methadone, methamphetamine/amphetamine
and other stimulants, narcotics, and sedatives.

DATA TYPE: survey data

TIME PERIOD: 1992 -1997

DATE OF COLLECTION: 1992 -1997

92

FUNDING AGENCY: United States Department of Health and Human Services.
Substance Abuse and Mental Health Services Administration. Center for Substance
Abuse Treatment.

GRANT NUMBER: 27 0-96-70 16

DATA SOURCE: personal interviews and patient records

DATA FORMAT: LRECL with SAS and SPSS data definition statements
COLLECTION NOTES:

(1) Data were collected by the National Opinion Research Center (NORC) at the
University of Chicago with assistance from Research Triangle Institute, Research
Triangle Park, NC. The NTIES public use files were prepared by NORC and deposited
with ICPSR by Caliber Associates, Arlington, VA, under the National Evaluation Data
Services (NEDS) contract with CSAT.

(2) To fully utilize the NTIES study design for analytic purposes, the combination of
records from two or more of the client-level data files is necessary. Some combination of
intake records (NRIQ), treatment exit records (NTEQ), treatment follow-up records
(NPAQ), treatment services abstracts (NPRAF), or clinical unit records (NCLU) is
needed to adequately assess changes in client behaviors over time. Client records can be
matched between data files using the CASEID variable.

(3) To protect the privacy of respondents, all variables that could be used to identify
individuals have been removed fi'om or edited in the public use file. Because of the
disclosure alterations, estimates derived from the public use file will not always exactly
match the detailed results published in SAMHSA reports, but the alterations should not

affect analytic uses of the data.

93

(4) Due to the disclosure alterations, there are slight inconsistencies in the categorical
ranges for "hours worked" variables between the NRIQ and NPAQ files. (5) Individuals
served by CSAT grants were generally from vulnerable and underserved populations
(minorities, pregnant women, youth, public housing residents, welfare recipients, and
those in the criminal justice system). Results from the NTIES may not generalize to all
clients in substance abuse treatment or to all kinds of service delivery units.

(6) The codebooks, data collection instruments, and frequencies are provided by ICPSR
as Portable Document Format (PDF) files. The PDF file format was developed by Adobe
Systems Incorporated and can be accessed using PDF reader software, such as the Adobe
Acrobat Reader. Information on how to obtain a copy of the Acrobat Reader is provided
on the ICPSR and SAMHSA Web sites.

SAMPLING:

(1) NTIES measured the outcomes of treatment primarily through a method known as a
"before/after" or "pre/post" panel design. From a universe of 698 SDUs, 82 SDUs were
selected on a purposive basis for participation in NTIES. Of the 82, clients from 78 SDUs
(for an SDU response rate of 95 percent) were included in the study. Clients were
interviewed three times: shortly after their first day of treatment, when they left
treatment, and then at 12 months after the end of treatment.

(2) The response rate among clients was 85 percent, with 6,593 clients participating in
the Intake Questionnaire (NRIQ), 5,274 participating in the Treatment Experience

Questionnaire (NTEQ), and 5,388 participating in the Post discharge Assessment

Questionnaire (NPAQ).

94

(3) Records abstraction was completed for 6,420 clients. The records of those
respondents participating in all three interviews are flagged by the variable "IN_4411" in
the NPAQ data file (Part 3). This is referred to as the Outcome Analysis Sample and
includes 4,411 records, or 67 percent of those participating in the initial interview. Some
cases were excluded from the analysis sample for reasons other than nonparticipation in
the three interviews, such as when the treatment exit date was missing or undetermined,
length of the interval for the follow-up interview was inappropriate (less than 5 or more
than 16 months), or the client was incarcerated for most or all of the follow-up period. Of
the SDUs sampled for the NTIES outcome analysis, 44 percent were Target Cities
programs, 38 percent were Critical Populations programs, and 23 percent were Criminal
Justice programs. Criminal Justice SDUs that were funded as part of the CSAT 1990-
1992 demonstrations were purposely over sampled as part of the NTIES evaluation
design. Nearly half of the sampled SDUs were non-methadone outpatient programs, and
about one-quarter were long-term residential programs.

UNIVERSE: Substance abuse treatment units in the United States receiving funding from
CSAT under one of three demonstration grants: Target Cities, Critical Populations, and

Criminal Justice.

95

APPENDIX B

NTIES Code Book Questions

96

NTIES Code Book Questions

Sample Questions from Original Survey

FILE SPECIFICATIONS

 

 

 

 

 

 

 

 

 

 

 

 

 

Part Part Name File Case Variable LRECL Records
No. i ‘ Structure Count Count Per Case
1 NTIES Research Intake rectangular 6,593 637 1,192 l
Questionnaire (NRIQ)

2 NTIES Treatment Experience rectangular 5,274 332 619 1
Questionnaire

3 NTIES Post-discharge rectangular 5,3 88 451 834 1
Assessment Questionnaire
(NPAQ)

4 NTIES Patient Record rectangular 6,420 224 526 l
Abstraction Form (NPRAF)

5 NTIES Clinical Unit Data rectangular 6,593 58 204 1
(NCLU)

 

Selected Questions from NRIQ Intake Questionnaire

 

LR29 GENDER

CODE GENDER AS OBSERVED. IF UNSURE, ASK:
I am required to ask if you are male or female. DO NOT CODE

REFUSED OR DON’T KNOW, CODE FROM ADMISSION FORM.

PCT PCT N
VALID ALL
72.1 72.1 4,753
27.9 27.9 1,840
100.0 100.0 6,593

Data type: numeric
Missing data codes: lowest thru — 1
Column: 97

VALUE

2

C8588

97

LABEL

MALE
FEMALE

 

 

[ R44M1 RACE/ETHNICIIY

What best describes you? Are you. . . .READ CATEGORIES.

PCT PCT N VALUE LABEL
VALID ALL
53.9 53.9 3,555 1 NONHISPANIC BLACK
14.8 14.8 974 2 HISPANIC
31.3 3 1.3 2,064 3 NONHISPANIC NONBLACK
100.0 100.0 6,593 cases

Data type: numeric
Missing data codes: lowest thru — 1
Column: 1 139

 

 

[ R58 HIGHEST GRADE ATTENDED

What is the highest grade or year of school you have ever attended,
even if you did not complete that grade?

PCT PCT N VALUE LABEL
VALID ALL
1.8 1.8 118 6 fimGradeorLowu'
1.8 1.8 117 7' 7" Grade
4.6 4.6 303 s 8" Grade
10.1 10.1 664 9 9“ Grade
16.6 16.6 1,097 10 10'“ crude
, 18.9 18.9 1,243 11 11“Grade
27.2 27.2 1,792 12 12" Grade
7.3 7.3 478 13 1 Year College/Technical School
6.8 6.8 448 14 2 Years College/Technical School
2.0 2.0 130 15 3 Years College/feclnucal School
3.1 , 3.1 203 16 4 Years College/Technical School
100.0 100.0 6,593 cases

Data type: numeric
Missing datacodes: lowwtthm—l
Column: 132 - 133

98

 

(R84M7 REFERRAL SOURCE — SPOUSE, PARTNER, FAMILY

Was a spouse, partner, or family member important in getting
you to come to (Note 1: treatment or counseling at this time/ this program)?

 

PCT PCT N VALUE LABEL
VALID ALL
35.3 35.3 2,327 1 YES
64.7 64.7 4,266 2 NO
100 0 100 0 6,593 cases
Data type: numeric
Missing data codes: lowest thru - 1
Column: 1118
[ R311 IMPORTANCE OF MENTAL HEALTH TREATMENT

 

Right now, how important to you is treatment for emotions, nerves, or
mental health problems? Would you say not at all important, somewhat important,

or very important?

PCT PCT
VALH) ALL
38.6 38.6
24.2 24.2
37.2 37.2
0.1

0.0

100.0 100 0

Data type: numeric

N VALUE
2,539 1
1,595 2
2,451

7 -2
l -1
6,593 cases

Missing data codes: lowest thru — 1

Column: 671-672

99

LABEL

NOT AT ALL IIVIPORTANT
SOMEWHAT MORTANT
VERY IMPORTANT
DON’T KNOW

REFUSAL

 

[R411 IMPORTANCE OF EMPLOYMENT COUNSELING

 

Right now, how important to you is counseling for employment problems, such as
problems finding or keeping a job, or getting along with the people you work with?
would you say not at all important, somewhat important or very important?

 

PCT PCT N VALUE LABEL
VALID ALL
38.4 38.3 2,527 1 NOT AT ALL IMPORTANT
19.6 19.6 1,293 2 SOMEWHAT IMPORTANT
42.0 42.0 2,768 3 VERY IMPORTANT
0.0 3 -2 DON’T KNOW
0.0 2 -l REFUSAL
100 0 100.0 6 593 cases
Data type: numeric
Missing data codes: lowest thru — 1
Column: 921 - 922
(R311 IMPORTANCE or HOUSING HELP

 

Right now, how important to you is help with housing problems such as finding or keeping
a place to live? ? Would you say not at all important, somewhat important,

or very important?

PCT PCT N
VALID ALL

49.4 49.4 2,539

10.7 10.6 702
40.0 40.0 2,634
0.0 3

100.0 100.0 6,593

Data type: numeric
Missing data codes: lowest thru — 1
Column: 447-448

VALUE

-2

C8588

100

LABEL

NOT AT ALL IMPORTANT
SOMEWHAT IMPORTANT
VERY IMPORTANT
DON’T KNOW

Selected Questions from NPAQ Post Discharge Assessment Questionnaire

 

 

[ P11 NTEQ COMPLETED?
PCT PCT N VALUE LABEL
VALID ALL
16.0 16.0 862 0 NO
84.0 84.0 4,526 1 YES
100.0 100.0 5,388 cases

Data type: numeric
Missing data codes: lowest thru — 1
Column: 57

 

[P172a3 DRUGS USED - CRACK

Show R Drug Cards, One at a Time, and Ask 172 for Each Drug Type:
Since {Note 1: STOP DATE/NTEQ DATE}, have you used crack, not
cocaine powder five times or more?

Note 1: See P138 Note 1

PCT PCT N VALUE LABEL
VALID ALL
19.7 19.7 1,061 1 YES, DRUGS USED
80.3 80.3 4,325 2 NO, DRUGS NOT USED
0.0 1 -2 DON’T KNOW
0.0 l -l REFUSAL
100.0 100.0 5,388 cases

Data type: numeric
Missing data codes: lowest thru — 1
Column: Ill-112

101

Selected Questions from NPRAF Patient Record Abstraction Form

 

[F15A3

#— RCVD: PSYCHIATRIST/PSYCHOLOGIST V.

 

15a3 Scan patient record for services received by patient during NTIES treatment
episode. For each service listed in Table 15, code in column A the number of times
the service was received. If the code in column A=8, go to next service. Then go to

next service.

PCT
VALID

58.8
17.6
7.6
4.1
11.9

100.0

PCT
ALL

12.6
3.8
1.6
0.9
2.6
0.0
1.8

76.8

100.0

Data type: numeric

Missing data codes: lowest thru - 1

Column: 278-279

809
242
105

VALUE

“AWNI—

cases

102

LABEL

ONE TIME

2 -— 3 TIMES

4 - 10 TIMES

11 OR MORE TIMES
NUMBER UNCERTAIN
NOT APPLICABLE
MISSING

NOT AVAILABLE

 

l F15A9 #- RCVD: EMPLOYMENT COUNSELING

15a9 Scan patient record for services received by patient during NTIES treatment episode.
For each service listed in Table 15, code in column A the number of times the service was
received. If the code in column A=8, go to next service. Then go to next service.

 

PCT PCT N VALUE LABEL
VALID ALL
12.9 2.9 188 1 ONE TIME
14.0 3.2 204 2 2-3 TIMES
13.4 3.1 196 3 4-10TIMES
39.1 8.9 572 4 11 OR MORE TIMES
20.7 4.7 302 7 NUMBER UNCERTAIN
0.0 1 -5 NOT APPLICABLE
1.1 73 -3 MISSING
76.1 4,884 -2 NOT AVAILABLE
100.0 100.0 6,420 cases
Data type: numeric
Missing data codes: lowest thru — 1
Column: 288-289
[P15A22 #- RCVD: HOUSING ASSISTANCE

 

15a22 Scan patient record for services received by patient during NTIES treatment episode.
For each service listed in Table 15, code in column A the number of times the service was
received If the code in column A=8, go to next service. Then go to next service

PCT
VALID

42.5
3.4
54.2

PCT
ALL

3.3
0.3
4.3
0.0

N

215

100.0

100.0

Data type: numeric
Missing data codes: lowest thru - 1
Column: 314-315

VALUE LABEL

103

1 ONE TIME
4 — 10 TIMES
7 NUMBER UNCERTAIN
-5 NOT APPLICABLE
-3 MISSING
-2 NOT AVAILABLE

cases

APPENDIX C

Type of Treatment Program

104

   
      
     
   

  

Type of
Treatment
Program

 

 

 

TX episode
completed

TX episode
completed

  

 

  

 

105

APPENDIX D

Variables Used for Analysis

106

Variables Used for Analysis

Variables used for analysis: Range

1:- Very Important 2=Somewhat Important 3 = Not at All Important
Treatment episode completed P11 1-2 yes no

Used Crack since Tx Completed P172A3 1-2 yes no
Treatment Modality R2 1-6”

Duration of Tx Episode LOS .25-42

Importance of housing help R190 1-3

Importance of child care counseling R248 1-4"I ‘Never raised children
Importance of family counseling R250 1-3

Importance of Social Problem

Counseling R251 1-3

Importance of mental health tx. R311 1-3

Importance of employment

Counseling R41 1 1-3

Importance of financial problems R420 1-3

Importance of medical problems R457 1-3

Need Child care R100M4 1-2 yes no

Need Transportation R100Ml 1-2 yes no

**1= Short term Hospital 3=Short term residential 4: Long term residential 5=Outpatient
methadone 6= Ambulatory outpatient

Received housing help F15A22 14
Received childcare counseling F15A15 1-4
Received family counseling F15A8 1-4
Received Social Problem

Counseling F15A14 14
Received mental health tx. F15A3 1-4
Received employment

Counseling F15A9 1-4
Received help financial problems F 15Al7 14
Received medical problems F15A2 1-4
Received Child care F15A21 1-2
Received Transportation F15A19 1-2

107

APPENDIX E

UCHRIS Letter of Approval

108

 

MICHIGAN STATE
u I R

N v E s 1 r Y Approval of
a public use
September 8, 2004 @jafig fiflg

To: Victor L. Whiteman
230 Baker Hall

Certification # PDO4-004

, Re: TREATMENT EXPERIENCES AND OUTCOMES FOR
AFRICAN AMERICAN WOMEN ADDICTED TO CRACK COCAINE

Thank you for submitting your registration for approval of a public
use data file. The University Committee-on Research Involving
Human Subjects (UCRIHS) has approved your request for the

use of Inter-University Consortium for Political and Social
Research (ICPSR) data file.

Please submit a Certification Form for the Use of an Approved
Public Data File for each specific project you wish to undertake
using this approved data file. This will allow you to conduct and
cOntinue each project without-IRB review. In addition, the
certification allows us to keep a total of all human research being
performed and gives credit to you and your department in any

' reports or statistics.

 

Thank you for your cooperation and good luck with your research.
_’ ' If we can be of further assistance, please contact us at 517-355-
OFF'CEOF - 2180 or via email at QQRIH§_@msu.edu.

 

RESEARCH
ETHICS AND Sincerely,
_ STANDARDS
University Committee on M g
Research Involving
Human Subjects
Michigan State University Peter Vasilenko, PhD
2020“” "a" UCRIHS Chair
East Lansing. MI
48824
517/355-2180 -
FAX: 517/432-4503 c: Janet Okagbue-Reaves
Web: www.msusdu/usei/ucrihs 640 Temple St.
E-Mail: ucrihs@msu.edu 6'" Fl

Detroit, MI 48201

109

MSU is an affirmative-action,
weal-opportunity institution.

REFERENCES

Bachman, R & Paternoster, R. (1997). Statistical methods for criminology and criminal
justice. New York, McGraw-Hill.

Bailey, K. D. (1994). Sociology and the new systems theory: Toward a theoretical
synthesis. Albany, NY, State University of New York Press.

Boyd, S. C. (1998). Mothers and illicit drugs: Transcending the myths. Toronto, Canada,
University of Toronto Press.

Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature
and design. Cambridge, Mass, Harvard University Press.

Buckley, W. F. (1967). Sociology and modern systems theory. Englewood Cliffs, N.J.,
Prentice-Hall.

Bureau of Justice Statistics, S. (1997). Center for Substance Abuse and Treatment,
National Treatment Improvement Evaluation Study 1997. Retrieved September 1,
2003, from http://www.health.org/nties97/costshtm.

Campbell, N. D. (2000). Using women: Gender, drug policy, and social justice. New
York, NY, Routledge Publishing.

Center for Substance Abuse and Treatment (1996). National Treatment Improvement
Evaluation Study. Washington DC, US Government Printing Office.

Center for Women Policy Studies (2003). Women, pregnancy, & substance abuse.
Report by the Gale Group.

Child Guide. (2003). Child Welfare Program Data Services Source. Retrieved February
2, 2003.
http://www.dss.cahwnet.gov/research/res/pdeChildGuide/CWSdatasource.pdf.

Colenso, H. M. (2002). Psychosocial correlates of drop out in a substance abuse
treatment program for mothers. Unpublished doctoral dissertation, School of
Social Work, New York University, New York, New York

Committee on Opportunities in Drug Abuse Research. (1996). Pathways of addiction

opportunities in drug abuse research. Washington, DC, National Academy
Press.

110

Dilonardo, J. (1998). The eflect of perceived self-relevance of wrap-around services on
treatment, retention and substance abuse outcome of adults. Unpublished doctoral
dissertation from, Human Development. University of Maryland, Baltimore

Finkbiner, R. (2000). Clients in treatment for alcohol problems: Treatment experiences
and outcomes. Journal of Psychopathology and Behavioral Assessment, Dec
2000.

French, M. T. (2002). Benefit—cost analysis of addiction treatment in Arkansas: Specialty

and standard residential programs for pregnant and parenting women. Substance
Abuse 23(1): 31-51(1).

Gerstein, D. R. (2000). Characteristics, services, and outcomes of treatment for women.
Journal of Psychopathology and Behavioral Assessment Dec 2000

Handel, W. H. (1993). Contemporary sociological theory. Englewood Cliffs, N.J.,
Prentice Hall.

Inciardi, J. A., Tims, F. M. (1993). Innovative approaches in the treatment of drug abuse:
program models and strategies. Westport, Conn., Greenwood Press.

Klee, H., M. Jackson, (2002). Drug misuse and motherhood. London; New York,
Routledge Publishing.

Lehmann, P. and N. Coady (2001). Theoretical perspecti ves for direct social work: A
generalist-eclectic approach. New York, Springer Pub.

Marilee Comfort, (2000). Predictors of treatment outcomes for substance-abusin g
women: A retrospective study. Substance Abuse 21(1): 33-45.

Maxfield, M. G. &. Babbie, E. R (2001). Research methods for criminal justice and
criminology. Belmont, CA, Wadsworth Pub.

McCoy-CB (2003). Welfare and work outcomes after substance abuse treatment. Social-
Service-Review 77(2): 237-254.

National Institute on Drug Abuse (2003). The economic costs of alcohol and drug abuse
in the United States. Retrieved September 21, 2003, fi'om the National Institutes
of Health, http://www.nida.nih.gov/EconomicCosts/Chapter1.html#1 .10

Nirenberg, T. D. and S. A. Maisto (1987). Developments in the assessment and treatment
of addictive behaviors. Norwood, N.J., Ablex Pub. Corp.

Office of Applied Studies (2002). National survey on drug use and health. Retrieved
March 5, 2003 fi'om the Substance Abuse and Mental Health Services

111

 

Administration.
httpzl/www.samhsa.gov/oas/nhsda/Zansduh/Overview/2k20verview.htm#chap2

Office of Applied Studies. (2002). Overview of findings fiom the 2002 national survey on
drug use and health. Retrieved November 11, 2003 from the Substance Abuse
and Mental Health Services Administration
httpzl/www.samhsa.gov/oas/nhsda/Zansduh/Overview/ZkZOverview.htm#chap8.

Office of National Drug Control Policy (2002). National Drug Control Strategy: FY 2003
Budget Summary. Washington, DC: Executive Office of the President: p. 6, Table
2.

Orwin, R. G. (2000). Relationships between treatment components, client-level factors,
and positive treatment outcomes. Journal of PsychOpathology and Behavioral
Assessment Dec 2000.

Patel, K. and M. E. Rushefsky (1999). Health care politics and policy in America.
Armonk, N.Y., ME. Sharpe.

Peele, S. (1989). Diseasing of America: Addiction treatment out of control. Lexington,
Mass, Lexington Books.

Plumb, M. (2000). Women and addiction. Retrieved from NIDA February 16, 2003
http://www.drugabuse.gov/PDF/DARHW/517-528_Plumb.pdf

Reif, S. (2002). The impact of ancillary services on substance abuse treatment outcomes.
Unpublished doctoral dissertation from The F. Heller Grad. Sch. for advanced
studies. Studies in Social Welfare, BRANDEIS University

Rios, R. (1994). Violence against women: The hidden epidemic. Washington DC. USA,
World Bank Discussion Paper 225.

Roberts, D. (2002). Race and the politics of child welfare. Retrieved February 2, 2003
from the Institute for Policy Research, Northwestern University.
http://www.northwestem.edu/ipr/publications/newsletter/imeOO6/roberts.htrnl

Sargaent, M. J. (1992). Women, drugs and policy in Sydney, London, and Amsterdam: A
feminist interpretation. Aldershot England; Brookfield, Vt., USA, Avebury.

Scott, L. (2000). The impact of women's family status on completion of substance abuse
treatment. The Journal of Behavioral Health Services & Research 27(4): 366-379.

Speck, K. J. (2002). A grounded theory of volition in recovery fi'om substance abuse.
Unpublished doctoral dissertation from The University of Nebraska, Lincoln.

112

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII

IIIIIIIIIILIIIIILIIIIIIIIIIIIII