A multidimensional approach to the study of prenatal cocaine exposure

MISSING IMAGE

Material Information

Title:
A multidimensional approach to the study of prenatal cocaine exposure
Physical Description:
xii, 115 leaves : ill. ; 29 cm.
Language:
English
Creator:
Edwards, Carla Denise Armbrister
Publication Date:

Subjects

Subjects / Keywords:
Sociology thesis, Ph. D   ( lcsh )
Dissertations, Academic -- Sociology -- UF   ( lcsh )
Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 2001.
Bibliography:
Includes bibliographical references (leaves 102-114).
General Note:
Printout.
General Note:
Vita.
Statement of Responsibility:
by Carla Denise Armbrister Edwards.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 028213447
oclc - 49320837
System ID:
AA00022656:00001

Table of Contents
    Title Page
        Page i
        Page ii
    Front Matter
        Page iii
    Acknowledgement
        Page iv
        Page v
    Table of Contents
        Page vi
        Page vii
        Page viii
    List of Tables
        Page ix
    List of Figures
        Page x
    Abstract
        Page xi
        Page xii
    Chapter 1. Introduction
        Page 1
        Page 2
        Page 3
        Page 4
        Page 5
        Page 6
    Chapter 2. The study of prenatal cocaine exposure
        Page 7
        Page 8
        Page 9
        Page 10
        Page 11
        Page 12
        Page 13
        Page 14
        Page 15
        Page 16
        Page 17
        Page 18
        Page 19
    Chapter 3. Multidimensional model
        Page 20
        Page 21
        Page 22
        Page 23
        Page 24
        Page 25
        Page 26
        Page 27
        Page 28
        Page 29
        Page 30
        Page 31
        Page 32
        Page 33
        Page 34
    Chapter 4. Research variables and sample description
        Page 35
        Page 36
        Page 37
        Page 38
        Page 39
        Page 40
        Page 41
        Page 42
        Page 43
        Page 44
        Page 45
        Page 46
        Page 47
        Page 48
        Page 49
        Page 50
        Page 51
        Page 52
        Page 53
        Page 54
        Page 55
        Page 56
        Page 57
        Page 58
        Page 59
    Chapter 5. Data analysis and results
        Page 60
        Page 61
        Page 62
        Page 63
        Page 64
        Page 65
        Page 66
        Page 67
        Page 68
        Page 69
        Page 70
        Page 71
        Page 72
        Page 73
        Page 74
        Page 75
        Page 76
        Page 77
        Page 78
        Page 79
        Page 80
        Page 81
        Page 82
        Page 83
        Page 84
        Page 85
        Page 86
        Page 87
    Chapter 6. Risk and resiliency in early childhood
        Page 88
        Page 89
        Page 90
        Page 91
        Page 92
        Page 93
        Page 94
        Page 95
        Page 96
        Page 97
        Page 98
        Page 99
        Page 100
        Page 101
    List of references
        Page 102
        Page 103
        Page 104
        Page 105
        Page 106
        Page 107
        Page 108
        Page 109
        Page 110
        Page 111
        Page 112
        Page 113
        Page 114
    Biographical sketch
        Page 115
        Page 116
        Page 117
Full Text











A MULTIDIMENSIONAL APPROACH TO THE STUDY OF PRENATAL COCAINE
EXPOSURE


B 4

CARLA DENISE ARMBRISTER EDWARDS












A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA


2001




























Copyright 2000

by

Carla Denise Armbrister Edwards





















This research was supported by two grants (5R01DA05854-08 and R2961444-12) from
the National Institute on Drug Abuse administered through the Department of Pediatrics
in the College of Medicine at the University of Florida. I extend special thanks to Fonda
Davis Eyler, Ph.D., the principal investigator and Marylou Behnke, the co-principal
investigator, for providing me with the opportunity to work with them on this project.















ACKNOWLEDGMENTS

It is difficult to know where to begin with my acknowledgements. I have been

blessed with a large and active support network. I am forever grateful to Dr. Barbara

Zsembik, my dissertation chair. She has provided me with guidance, motivation, and

wisdom. I am especially thankful to Dr. Leonard Beeghley, Dr. Terry Mills, and Dr.

John Henretta, my committee members. I truly appreciate how they always required

more and how they anticipated the best from me at all times. I express sincere

appreciation to Dr. Fonda Davis Eyler and Dr. Marylou Behnke for inviting me to serve

on their research team, for mentoring me, and for providing me with numerous

opportunities to enhance my research skills and experience.

Also, I must acknowledge the women and children who have permitted the

research team to delve into their lives over the past ten years. I hope the results of this

research will serve to enhance the quality of life for the many women and children who

are struggling to cope with their mere existence on this planet.

This dissertation would not have come to fruition without the love and support of

my husband, Herman Edwards and my parents, Anthony and Juanita Armbrister. Over

the past four years, my immediate and extended family has encouraged me to finish my

degree while helping me cope with the loss of many loved ones (including the loss of my

only sibling). The encouragement, love, and support I have received from this large

group of people, including two grandmothers, aunts, uncles, and cousins, is

immeasurable.








Last, I am thankful for my Aunt Priscilla, who read my manuscript, my Uncle

Rickie, who hugged me and pointed me in the right direction after I lost my manuscript,

my dear friend Yvonne, who helped me find it again, and my Aunt Lucille, who delivered

it to the graduate school for first submission. Along with several other angels, these folks

were all in the right place at the right time.
















TABLE OF CONTENTS

page

ACKN OW LED GM ENTS.................................................................................................. iv

LIST OF TABLES ............................................................................................................. ix

LIST O F FIGURES............................................................................................................. x

ABSTRACT ....................................................................................................................... xi

CHAPTERS

1 INTRO DUCTION ........................................................................................................... 1

2 THE STUDY OF PRENATAL COCAINE EXPOSURE.............................................. 7

Literature On Prenatal Cocaine Exposure....................................................................... 7
Biom medical Research................................................................................................... 9
Behavioral (Psychological) Studies .......................................................................... 13
Sociological Studies ............................................................................................... 17
Conclusion ..................................................................................................................... 18


3 M ULTID IM ENSION AL M OD EL ..................................................... ..................... 20

Sociological Perspective ............................................................................................... 21
Com ponents of the M ultidim ensional M odel ............................................................... 22
Prenatal Dim ension ................................................................................................... 24
Neonatal D im ension .................................................................................................. 24
Postnatal D im ension.................................................................................................. 25
Theoretical Framework ................................................................................................. 25
Social Ecological Theory .......................................................................................... 25
Resiliency Theory ..................................................................................................... 29
Definitions O f Behavior And Developm ent.................................................................. 31
Reason for Studying Pre-Schoolers........................................................................... 32
Research Q questions ....................................................................................................... 33










4 RESEARCH VARIABLES AND SAMPLE DESCRIPTION..................................... 35

Data Collection .............................................................................................................. 35
Description of Longitudinal Sam ple and Sam ple Retention......................................... 37
Description Of Birth M others ...................................................................................37
Characteristics O f Children At Birth......................................................................... 39
Retention And Attrition At Three Year Follow Up .................................................. 42
Operationalization Of V ariables and H ypotheses......................................................... 43
Prenatal Risk: Prenatal Cocaine Exposure (PCE).................................................... 43
O utcom e M measures: Behavior and Developm ent...................................................... 46
Neonatal Risk: N ewborn Health Status..................................................................... 48
Postnatal Factors: Social Ecological M odel.............................................................. 49
Sum m ary ................................................................................................................... 57


5 DATA AN ALYSIS AND RESULTS........................................................................... 60

Analytical Procedures ..................................................................................................60
Differences between PCE children and non-PCE children........................................... 62
Birth Characteristics....................................................... ........................................ 62
Neonatal Environm ent......................................................... ...................................63
Postnatal Environm ent .............................................................................................. 65
Behavior and D evelopm ent....................................................................................... 70
Environm ental Factors and Early Childhood O utcom es............................................... 73
Prenatal Risk and O utcom es ..................................................................................... 73
Neonatal Risk and Outcom es .................................................................................... 73
Postnatal Factors and O utcom es ............................................................................... 75
M ultidim ensional M odel of Behavior and Developm ent.............................................. 80
Behavior .................................................................................................................... 81
Develop ent ............................................................................................................. 83
Sum m ary ....................................................................................................................... 86


6 RISK AND RESILIENCY IN EARLY CHILDHOOD ............................................... 88

Early Childhood Risk.................................................................................................... 88
Prenatal Cocaine Exposure........................................................................................ 88
Poor Neonatal Health ................................................................................................ 91
M aternal D expression ................................................................................................. 92
Low Socioeconom ic Status ....................................................................................... 94
Producers Of Resilience................................................................................................ 96
Positive Parenting...................................................................................................... 96
Fam ily and Social Support........................................................................................ 97
Interventions And Im plications..................................................................................... 98
Lim stations And Recom m endations......................................................................... 100










LIST OF REFEREN CES ................................................................................................ 102

BIO GRA PH ICAL SKETCH ........................................................................................... 115















LIST OF TABLES


Table Page

4. 1: Descriptive Statistics of Biological Mother and Child at Birth by Cocaine Exposure
Status, Means +/- Standard Deviations (Mean Rank)......................................... 39

4. 2: Longitudinal Study Sample Birth Outcomes by Cocaine Exposure Status, Means +/-
Standard Deviations (M ean Rank)...................................................................... 40

4. 3: Subject Retention and Attrition at 3-Year Follow Up ................................................. 43

4. 4: Biological Mothers Drug Use by Cocaine Exposure Status, n=308............................ 45

4. 5: Caregiver's Relationship to Child at Age Three by Cocaine Exposure Status............. 50

4. 6: Description of Variables in Multidimensional Model .................................................. 59

5. 1: Birth Outcomes by Cocaine Exposure Status, Means+/-Standard Deviations (Mean
R an k ) ................................................................................................................... 6 2

5. 2: Postnatal Differences between PCE and non-PCE children at the Microsystem and
Mesosystem Level, Mean +/- Standard Deviation (Mean Rank)........................ 67

5. 3: Postnatal Differences between PCE and non-PCE children at the Exosystem and
Macrosystem Level, Mean +/-Standard Deviation (Mean Rank)....................... 69

5. 4: Adaptive Behavior and Development by Cocaine Exposure Status, Means +/-
Standard Deviations (M ean Rank)...................................................................... 72

5. 5: Correlation between Adaptive Behavior and Development, Prenatal Cocaine
Exposure, and Environmental Factors, Spearman Correlation Coefficient,
n = 2 34 ................................................................................................................... 76

5. 6: Complete Correlation Matrix, Spearman Correlation Coefficients, n=234................. 79

5. 7: Multidimensional Models of Adaptive Behavior and Development, Standardized
Regression Coefficients (p-values), n=234......................................................... 83





















LIST OF FIGURES


Figure Page

2. 1: Number of Citations per Discipline Per Year............................................................... 9

3. 1: Compensatory and Protective Models of Resilience .................................................... 29

3. 2: Multidimensional Model of Behavior and Development.............................................. 31

4. 1: Percent of Low Birth Weight Children by Prenatal Cocaine Exposure Status, n=302.41

4. 2: Percent of Weeks Cocaine Used During Pregnancy, n=1 15......................................... 45

5. 1: Hobel Complications Scale by Prenatal Cocaine Exposure Status............................... 65















Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

A MULTIDIMENSIONAL APPROACH TO THE STUDY OF PRENATAL COCAINE
EXPOSURE
By

Carla Denise Armbrister Edwards

December 2001


Chairman: Barbara A. Zsembik, Ph.D.
Major Department: Sociology

OBJECTIVE: Using longitudinal data this research seeks to determine the

prenatal, neonatal, and postnatal predictors of adaptive behavior and early childhood

development among children who were prenatally exposed to cocaine. This research

uses a multidimensional and integrative approach to examine how sociological,

behavioral, and biomedical factors influence the behavior and development of a group of

children "at-risk" of poor developmental outcomes. The respective influence of prenatal

cocaine exposure (PCE), the health status of the infant at birth, the psychosocial status of

the primary caregiver, the home environment, the families' social supports, and social

structural variables are included as determinants.

METHOD: This study utilizes data collected for a longitudinal study on the

impact of prenatal cocaine exposure on children. The study sample includes 115

prenatally cocaine exposed children and 119 controls. The sample is predominantly

comprised of poor, Black children from single parent households who are "at-risk" of








developmental delays because of the problems associated with maternal drug use, a

typical consequence of social and economic disadvantage. Using bivariate and

multivariate analysis, this study tests the utility of a multidimensional approach designed

to contribute to our understanding of the relationship between PCE and the measured

outcomes.

RESULTS: The bivariate and multivariate analyses indicate that there is no

relationship between PCE and the behavioral and developmental outcomes of three-year

old children. Both behavior and development are significantly correlated with the home

environment, the families' social support mechanism, and the families' social

demographics. The quality of the parenting environment within the home is the strongest

predictor of behavior followed by the families' social demographics and the infant's

health status at birth. The infant's neonatal health status and the families' social

demographic were the most significant predictors of development. The multidimensional

model, which integrates measures of the prenatal, neonatal, and postnatal environments,

proved to be a stronger model than the prenatal or prenatal and neonatal model combined,

thus providing evidence that despite the risk associated with prenatal cocaine exposure,

"at-risk" children exhibit resiliency when placed in positive postnatal environments.














CHAPTER 1
INTRODUCTION

Children who have been prenatally exposed to crack cocaine have been

characterized as "at-risk" of exhibiting a variety of developmental and behavioral

problems (Azuma and Chasnoff 1993; Chasnoff 1988; Chasnoff, Lewis, Griffith, Willey

1989; Frank, Zuckerman, Amaro, Aboagye, Bauchner, Cabral, Fried, Hingson, Lynne,

and Levenson 1988; Tronick, Frank, Cabral, Mirochnick, and Zuckerman 1996).

According to the Robert Wood Johnson Foundation, over 80,000 children will have

subtle deficits in their intellectual and language development as a consequence of

prenatal cocaine exposure. These presumed deficits have an anticipated cost to the nation

of nearly $352 million per year for special education programs (Lester, LaGasse, and

Seifer 1998). It has not been determined through the use of sound empirical research

methods whether or not the anticipated and observed problems are a consequence of the

actual physiological effect cocaine has on the developing fetus and surviving child or the

social and environmental consequences of being born to a drug addicted mother who may

have difficulty caring for her child.

There is a long history of research on the numerous social and environmental

factors that contribute to children's delayed development and problematic behavior. In

many of these studies, African-American children are disproportionately considered "at-

risk" due to the adverse effects of poverty on their health and over all well-being (Federal

Interagency Forum of Child and Family Statistics 1997; U. S. Department of Health and








Human Services 1995). Poor children are "at-risk" because of the problems associated

with family and community disintegration (Kendall-Tackett and Eckenrode 1996; Wilson

1987) and family violence (Deater-Deckard, Bates, Dodge, and Pettit 1996; Dodge, Pettit,

Bates 1994; Kendall-Tackett and Eckenrode 1996). Disconcertingly, poverty, family and

community disintegration, and poor family dynamics are factors that characterize the

lives of women who abuse drugs and the lives of the children they bear (Coy 1997).

Consequently, it is difficult to determine whether or not prenatal cocaine exposure or

some postnatal social environmental factor is responsible for the developmental and

behavioral problems presumably exhibited by "at-risk" children.

This study uses an integrative approach to examine the sociological and the

biomedical determinants of early childhood behavior and development for children

defined as "at-risk" due to their prenatal cocaine exposure. Specifically, this research

will address the following: (1) the differences in the behavior and development of

prenatally and non-prenatally cocaine exposed children and (2) to what extent the

behavior and development of pre-school aged children is either enhanced or hindered by

biomedical, behavioral, and sociological factors found in the prenatal and postnatal

environments.

Drawing from the health integrative approach advocated for by the Committee on

Future Directions for Behavioral and Social Sciences Research at the National Institutes

of Health (Singer and Ryff 2001), I hypothesize that the behavior and development of

young children are not only dependent on physiological (biomedical) factors derived in

the prenatal environment, but that they are highly dependent on neonatal factors, such as

the child's health status at birth and postnatal factors, such as the social structure, social








support networks, the physical environment, and interpersonal relationships. This study

seeks to determine the influence of prenatal, neonatal, and postnatal factors on the

behavior and development of children at age three, a critical stage of maturation in early

childhood development.

The predominant belief upheld by popular culture is that children who have been

prenatally exposed to cocaine are "at-risk" of poor developmental outcomes and that their

behavioral problems are placing a drain on social, educational, and economic resources

(Blakeslee 1990; Bragg 1998; Gross 1993; Hinds 1990; Toufexis 1990). The media have

helped perpetuate this common belief by creating the image of the "crack baby" and the

out of control drug exposed child or "crack kid." However, much of the scientific

research on prenatally cocaine exposed children has provided very little evidence that

cocaine exposed children have any more problems than other children who have survived

unhealthy prenatal experiences.

Much of the research on the behavior and development of young children who

have been prenatally exposed to cocaine has been contradictory and/or inconclusive.

There are several reasons for the inability of researchers to establish a clear connection

between prenatal cocaine exposure and subsequent behavior and development. One

reason is that there are only a few longitudinal studies on prenatally cocaine exposed

children. Moreover, most researchers are still in the process of collecting and analyzing

data that include outcome measures for pre-school and school-aged children. Second, the

majority of studies that have been published have taken a biomedical or behavioral

approach ignoring the role of social and environmental factors.








This study is empirically unique in that the data come from a longitudinal project

funded by the National Institute on Drug Abuse, which contains data for a matched

control group of prenatally cocaine exposed children from birth through age seven. This

research addresses the developmental concerns of the children at age three. Furthermore,

this research is theoretically unique, because it is written from the perspective of a social

scientist with a multidisciplinary background including training in sociology,

neuropsychology, counseling, and education. This background predisposes me to

integrate multiple disciplinary ideas, as I seek to understand how numerous systems

interact with one another and how these systems influence different aspects of our social

lives.

In this research, I raise the question of whether or not prenatal cocaine exposure, a

consequence of a mother's behavior during pregnancy, is more influential than the

biomedical health risk determined neonatally and postnatal factors, including

socioeconomic status, family social support, the home environment, and the primary

caregiver's psychosocial status. As a social science researcher, I intend to shed light on

a topic that has been dominated by medical and behavioral researchers. By identifying

the additive effect of biomedical and social environmental factors associated with

behavioral and developmental outcomes among "at-risk" children, researchers and

practitioners can better ascertain the role these factors play in the lives of prenatally

cocaine exposed children. Then, we can create interventions that may reduce the

negative impact of the prenatal environment by accentuating the positive attributes found

in the postnatal environment.








Using a sociological perspective and the basic tenets of human ecology and

resiliency theory outlined in Chapter 3, I have created a conceptual model that will be

used to ascertain the effect of the prenatal, neonatal, and postnatal environments on the

behavior and development of the children included in this research. The compensatory

model of resilience describes the mechanism by which the neonatal and postnatal

environments mediate the effect of the prenatal environment on early childhood

outcomes. The children's prenatal exposure to cocaine represents the prenatal influence

and the first layer of influence over behavior and development. The neonatal dimension

is characterized by the newborn's risk of experiencing medically related problems and it

is related to the outcome. The postnatal dimension is comprised of the social ecological

systems described in Bronfenbrenner's (1979) ecological model of human development.

These systems are characterized by the following: (1) the families' sociodemographic

background, (2) the families' social support system, (3) the home environment, and (4)

the psychosocial status of the primary caregivers.

The data in this research come from a matched-control prospective, longitudinal

study on the effects of prenatal cocaine exposure on fetuses, infants, and developing

children. The study includes a target group of children whose biological mothers used

cocaine during their pregnancy and a control group of children whose biological mothers

did not use cocaine during their pregnancy. During the enrollment that began during the

pregnancy, for each cocaine user enrolled, a non-cocaine user was enrolled who had

similar sociodemographic characteristics and prenatal health risks.

The strengths of the data include the following: (1) the early prenatal enrollment

of cocaine users and non-users who were matched on numerous social and physiological








demographics, (2) the repeated drug measures, including oral maternal drug use histories,

unexpected urine screenings, and confirmation of cocaine exposure through gas

chromatography/mass spectroscopy, and (3) the use of well trained evaluators who did

not know the children's cocaine exposure status while conducting the child assessments

of behavior and development.

The goal of this research is to examine the problems associated with prenatal

cocaine exposure using a multidisciplinary approach and an empirically sound research

design that attempts to avoid the pitfalls found in the existing research on the topic of

behavior and development of "at-risk" children. The proceeding sections of this report

include a summary of the existing literature on prenatal cocaine exposure (chapter 2), an

outline of the multidimensional model used to evaluate behavior and development

(chapter 3), a detailed description of the research design and sampling procedures

(chapter 4) and an analysis of the findings (chapter 5) along with their practical and

theoretical implications (chapter 6).














CHAPTER 2
THE STUDY OF PRENATAL COCAINE EXPOSURE


The image of a drug-exposed infant lying in an incubator connected to lifesaving

machinery has been reproduced in magazines, newspapers, journals, and on television

throughout the 1980s and 1990s. The rising number of pregnant cocaine users giving

birth to prenatally exposed infants and the emotions provoked by these media drawn

images resulted in the emergence of over 100 scientific studies on the effects of prenatal

cocaine exposure on women and children (Lester, LaGasse, and Seifer 1998). Based on

my review of the National Institutes of Health's listing of scientific projects, there are

over thirty federally funded studies on prenatal cocaine exposure (National Institutes of

Health 2001).

In this chapter, I summarize the biomedical, behavioral, and sociological literature

on prenatal cocaine exposure. The literature that I review denotes the need for more

studies on prenatally cocaine exposed children that take an integrative or

multidisciplinary approach. There is an abundance of research on the physiological

effects of cocaine on fetuses and newborns and a paucity of research on the social

environmental determinants of early childhood development for this group of children.


Literature On Prenatal Cocaine Exposure

My review of the literature on prenatal cocaine exposure consisted of a search of

library databases using the terms "prenatal cocaine exposure" and/or "cocaine and

pregnancy." This search resulted in over 800 articles that I have placed in four








categories: biomedical studies, behavioral (psychological) studies, sociological research,

and media generated articles.

I was able to find research articles in (1) Medline, the most popular search engine

for medically related scientific research, (2) Psyclnfo, a search engine including articles

related to psychology and the behavioral sciences, (3) Sociological Abstracts and

Sociofile, two search engines dedicated to social science research, and (4) Lexis-Nexis, a

search engine that includes international and national newspaper and magazine articles, in

addition to other sources of information. Due to the nature of computer databases, the

overlap between databases is inevitable. There were articles in Medline that were also

available in Psyclnfo and research in Psyclnfo and Medline that may have been covered

by newspapers and magazines listed in Lexis-Nexis, but this does not lessen the point that

there were more articles on prenatal cocaine exposure and cocaine use in pregnancy in

the biomedical and psychological databases than in the sociological databases.

There were 244 scientific articles in Medline, characterizing a plethora of research

in the biomedical field. There were 235 articles posted in Psyclnfo compared to only 20

articles (less than 5%) in the sociological databases, Sociological Abstracts and Sociofile.

Thus, nearly half of the articles were found in the biomedical and psychologically

oriented databases with the remaining 40% (n=343) being media generated articles found

in Lexis-Nexis. In the academic disciplines, the number of articles produced between

1985 and the year 2000 increased substantially. Between 1990 and 1995 the media's

interest in the topic reached its peak at 174 articles, outnumbering the articles found in

Medline (n=101), Psyclnfo (n=110), and the sociological databases (n=9). Figure 1

illustrates the distribution of the literature across each area of interest.




















500-1
400-
300- "- "
200-
100-
0'
1985-90 1990-95 1996-2000 Total
MMedline 17 101 126 244
SPsycinfo 25 100 110 235
-Socabs 1 9 10 20
OMedia 83 174 86 343
* Total 125 384 332 842


*Medline

* Psycinfo

OlSocabs

0 Media

MTotal


Figure 2.1: Number of Citations per Discipline Per Year



Biomedical Research

When the "crack scare" began in 1985, the majority of research on cocaine

exposed newborns focused on the potential neurological and physiological damage

caused by the drug using a teratogenic or deficit model approach. Health care providers,

researchers, educators, and social servants developed a deep concern for prenatally

cocaine exposed children as a consequence of these early reports of severe prenatal and

postnatal problems caused by pregnant women's use of crack cocaine. As seen in media

reports on the prevalence of the problem, many people also became anxious about the

impact prenatally cocaine exposed children would have on our health and social service








systems (Archibald 1992; Bays 1990; Martinez 1995). These concerns resulted in a

drastic increase in the number of research studies on the topic. However, many of these

studies lacked either an inclusive theoretical perspective or a sound methodological

framework (Zuckerman and Frank, 1994).

Early biomedical studies reported problems such as spontaneous abortions

(Cherukuri, Minkoff, Feldman, Parekh, and Glass 1988; Lutiger, Graham, Einarson, and

Koren 1991; Ryan, Ehrlich, and Finnegan 1987; Wang and Schnoll 1987), premature

births (Cherukuri et al. 1988; Rosenak, Diamant, Yaffe, and Hornstein, 1990), low birth

weight (Cherukuri et al. 1988; Chouteau, Namerow, and Leppert 1988; Ryan et al. 1987;

Zuckerman et al. 1989), and irreversible birth defects (Chavez, Mulinare, and Cordero

1989) in cocaine exposed human newborns. Yet, more recent studies have published

contradictory evidence due to differences in their study design and use of different

statistical controls.

For example, both Cherukuri et al. (1988) and Chouteau et al. (1988) reported that

cocaine use during pregnancy caused low birth weight among exposed infants, but they

both did not find similar results when comparing the relationship between prenatal

cocaine exposure and spontaneous abortions or premature births. In the Cherukuri et

al.(1988) study, which did not control for the use of prenatal care, cocaine exposure was

related to preterm births, whereas in the Chouteau et al. (1988) study, which controlled

for prenatal care and a variety of social demographic factors, cocaine exposure was not

related to pre-term births. According to Chouteau et al. (1988), lack of prenatal care was

a significant predictor of this poor health outcome.








The following studies published in the late 1980s also failed to control for social

and environmental factors that could substantially influence the study results. In 1989,

Zuckerman and his associates reported the findings from a prospective study that

included 1226 pregnant women and their newborn infants. In their study on the impact of

cocaine exposure on fetal outcomes, the researchers found that when comparing the

newborns of women who tested positive for cocaine (n=l 10) to non-cocaine users

(n=l 116), the newborns of cocaine users had significantly lower birth weights (p = 0.07),

a decrement in length (p = 0.01), and proportionally smaller head circumferences (p =

0.01) than the non-exposed newborns (Zuckerman et al. 1989). Even though the number

of the cocaine-exposed infants included in the study was considerably smaller than the

non-cocaine exposed infants, the researchers were able to control for a variety of

physiological factors in this study; yet, social and environmental factors that can

influence neonatal outcomes were not controlled.

Chavez, Mulinare, and Cordero (1989) conducted a retrospective study of

congenital anomalies in children born in Atlanta area hospitals between 1968 and 1980.

In this study, the researchers report a statistically significant relationship between

maternal cocaine use and urinary tract anomalies in infants (crude odds ratio, 4.39; 95%

confidence interval, 1.12 to 17.24). However, this study did not control for the social,

demographic, or environmental factors that may also relate to poor birth outcomes, such

as race, access to prenatal care or socioeconomic status.

Kliegman and associates (1994) conducted a prospective study of the effect of

prenatal cocaine exposure on birth weight and prematurity. In the Kliegman (1994)

study, of the eleven sociodemographic and physiological variables analyzed, race and








utilization of prenatal care services, along with cocaine, marijuana, and cigarette use,

were significantly correlated with poor birth outcomes. Cocaine exposure (odds ratio,

13.4; 95% confidence intervals, 1.23 to 145.0) and prenatal care (odds ratio, 9.89; 95%

confidence intervals, 3.74 to 26.17) were significant predictors of low birth weight and

premature births, demonstrating a need to assess the long-term effects of these variables

on developmental outcomes.

Most studies concur that prenatal cocaine exposure results in low birth weight and

prematurity, but there are also other factors, like the lack of prenatal care, race, and

overall maternal health, that are related to these poor health outcomes. In 1997, Lester

and his associates constructed a database of studies on prenatal cocaine exposure (Lester,

LaGasse, and Brunner 1997). The 76 studies included in the final database (1) pertained

to cocaine use during pregnancy, (2) were based on original research using human

subjects, (3) included neurobehavioral measures, (4) included control or comparison

groups, (5) analyzed the data statistically, and (6) were found in peer reviewed or

refereed publications.

The analysis of the articles included in the database lead Lester and his colleagues

to contend that the neurodevelopmental disorders identified in infants participating in the

studies on prenatal cocaine exposure could be attributed only partially to the presumed

neurological damage caused by drugs. Lester et al. (1997:487) writes:

The data base shows that our knowledge base is limited, scattered, and
compromised by methodological problems that mitigate any conclusions
about whether or not or how prenatal cocaine exposure affects child
outcome. Only a few studies have followed children beyond age 3. In
addition, the cocaine problem is more complicated than first envisioned.
It is a multifactorial problem including the use of other drugs and
parenting and environmental lifestyle issues.








Lester and his colleagues (1996, 1997) along with other behavioral researchers

(Graham and Koren 1991) have suggested that there are social, environmental, and other

medical conditions that may contribute to the poor health outcomes of prenatally exposed

children, but as indicated by my review of the literature and Lester's database, very little

of the research conducted by natural scientists address social concerns. Of the articles I

reviewed, none of the biomedical studies were initially designed to address the social and

environmental concerns that many behavioral and social science researchers deem

important.

Behavioral (Psychological) Studies

The psychological and behavioral research that addresses the social and

environmental conditions affecting the development of prenatally exposed newborns has

helped shift the focus from the biomedical to the social and environmental. Many of the

psychologically oriented studies of prenatal cocaine exposure focus on how cocaine

exposure influences the behavioral development of newborns and young children. This

differs from the biomedical focus on the effects of cocaine on human physiology and

subsequent health outcomes. The behavioral studies also tend to include more

psychosocial controls, thus providing further explanation of the differences or lack of

differences between children who have been prenatally exposed to cocaine and non-

cocaine exposed children.

For example, Woods and associates conducted a study of the effect of cocaine

exposure and maternal affect on the neurobehavioral development of neonates (Woods,

Eyler, Behnke, and Conlon 1993). This study included maternal depression and prenatal

care as psychosocial controls. In this well-controlled and methodologically sound study,

the researchers did not find any significant differences in the neurobehavioral








development of one-month old infants who were and who were not exposed to cocaine

prenatally. This study from the larger research project from which I am drawing my data,

employed a theoretical framework that acknowledged the importance of psychosocial

factors, while using a sound quasi-experimental design.

Many of the studies found under the behavioral or psychological rubric have

employed similar experimental designs that are intended to control for physiological and

environmental factors that may influence the results. Some of the researchers have

acknowledged the influence that the social environment has on developmental outcomes

(Bendersky and Lewis 1998; Grant, Bendersky, and Lewis 2000; Alessandri, Bendersky,

and Lewis 1998), but there are still many studies that lack sound methodological

guidelines making it difficult to present conclusive results.

In a recent article by Singer (1999) titled "Advances and Redirections in

Understanding Effects of Fetal Drug Exposure," the author calls for researchers to use

more "rigor" in their methodological designs, to improve "quantifiable and self report

measures," and to use more "sensitive" statistical models. These comments are a

response to the inconclusiveness of the existing research on children who have been

prenatally exposed to a variety of drugs, including cocaine.

Many biological and behavioral researchers have been unable to conclusively

state that cocaine use during pregnancy directly impacts later behavior in the pre-exposed

child. The studies that have been published utilize longitudinal, cross-sectional,

retrospective, prospective, and case controlled research designs and they vary in how

prenatal cocaine use is measured, defined, and analyzed (Mayes, Granger, Bomrnstein, and

Zuckerman 1992). The inconsistency in research methodologies makes it very difficult








to summarize the research findings pertaining to the effect of prenatal cocaine exposure

on behavior and development from a biomedical or a behavioral perspective. Linda

Mayes' (1996) review of the literature found that most studies dealing with behavioral

and developmental problems among cocaine exposed children were plagued with

methodological problems, such as poor population sampling, issues of polydrug use, lack

of longitudinal data, and lack of environmental controls.

For example, the variability in research designs has lead to contradictory results in

several behavioral and developmental studies. In an article written by Angelilli and

associates (1994), it is stated that children with a detected language delay are more likely

to have been prenatally exposed to cocaine than children with normal language

development. However, this finding is based on a study of language delay in which 29

children with delays were compared to children with no delays and prenatal drug use was

collected retrospectively from the child's caregiver during the child's clinical

appointment. The researchers found prenatal cocaine and nicotine exposure to be

significantly correlated with language delays among children aged 24 to 48 months. The

research design does not allow for adequate controls nor does it verify the nature of the

cocaine exposure.

In a prospective longitudinal study of prenatal cocaine exposure with masked

evaluators conducted by Hurt and her colleagues (1997), the researchers did not find any

relationship between poor language development and prenatal cocaine exposure among

their matched sample of prenatally cocaine and non-cocaine exposed subjects at 30

months of age using the Preschool Language Scale. However, in a different study,

Johnson and his peers (1997) found significant differences in language ability and








development skills between prenatally cocaine exposed and non-exposed children

between 11 and 50 months using the Sequenced Inventory of Communicative

Development-Revised and the Bayley Scales of Infant Development. This study included

a matched sample of children who had been living in stable, drug free environments at

the time of the testing.

Eyler and Behnke (1999) created a comprehensive review of the literature on

prenatal cocaine exposure's influence on early development. Like Mayes (1996), Eyler

and Behnke also comment on the variability in methodologies that make it difficult to

make conclusive statements about cocaine's influence on early childhood development.

According to Eyler and Behnke (1999), approximately one third of the neonatal studies

they reviewed found that prenatal cocaine exposure did not influence behavioral

development. The most commonly reported findings were poorer state and autonomic

regulation, irritability, and decreased alertness and orientation among cocaine exposed

neonates. In studies of three to twenty-four month old infants, about half found no

differences between prenatally cocaine exposed and non-exposed infants in development

or adaptive behavior. According to Eyler and Behnke (1999), other studies found that

prenatally cocaine exposed children experienced problems with arousal and

responsiveness, visual expectancy, recognition memory, information processing, and

attachment and play during the first year of life.

Singer (1999) proposes more systematic questioning and longitudinal research as

a means of determining the long-term effects of cocaine on behavior, while I propose

different questions and more sociological research as a means of increasing our








understanding of the relationship between prenatal cocaine exposure and childhood

outcomes.

Sociological Studies

In my review of the literature, I was able to identify only 20 studies on "prenatal

cocaine exposure" and/or "cocaine and pregnancy" in the sociology databases. Only

four of these articles pertained to prenatally cocaine exposed infants. The majority of the

articles were about drug use among pregnant women (Bendersky, Alessandri, Gilbert,

and Lewis 1996; Choe, Murphy, and Murphy 1998; Higgs 1996; Humphries 1998;

Humphries, Dawson, Cronin, Keating, Wisniewski, and Eichfeld 1992; Kearney and

Murphy 1993; Kearney, Murphy and Rosenbaum 1994; Lanehart, Clark, Kratochvil,

Plings, and Fidora 1994; Maher 1990; Murphy and Ferreboeuf 1997; Pursley-Crotteau

and Stern 1996; Singer, Arendt, Minnes, Farkas, Yamashita, and Kliegman 1995; Smith,

Dent, Coles, and Falek 1992; Teagle and Brandis 1998). The articles that dealt with

prenatally cocaine exposed infants addressed issues of care-giving (Franck 1996; Hamid

1994), the protection of prenatally exposed infants (Besharov 1989), and the problems

these children present to society (Best 1994). Unlike the biomedical and behavioral

studies, these studies did not focus on the effect prenatal cocaine exposure has on the

developing child, but rather the impact prenatally exposed children have on the world in

which they live.

There are numerous methodological and theoretical gaps within the sociological

literature, given the limited number of articles and their limited scope. There are studies

that address social issues, such as the relationship between prenatal cocaine exposure and

the parenting, home, and community environments, but these were not found within the

sociological databases used in this literature review, because these studies lacked a clear








sociological perspective. Therefore, my research is intended to fill the gap by conducting

research on prenatal cocaine exposure from a viewpoint that is not only sociological, but

multidimensional in nature as well.

In any research that depends on the study of human subjects within their social

environment, it is very difficult to maintain a strict disciplinary focus while conducting

research that requires a rigorous experimental design. Human subjects are difficult to

study because what happens in the course of the human subjects' life when outside of a

laboratory is impossible to control. Unlike rats used in experimental research, it is nearly

impossible and in virtually all cases unethical to place humans in a box or a controlled

environment for a prolonged period of time to analyze their behavior while exposing

them to a lethal substance, such as cocaine. Social science researchers can benefit from

the methods used by behavioral scientists, and biomedical scientists can benefit from the

social context incorporated into research studies by social scientists. Hence, a

multidimensional integrative approach needs to be employed.


Conclusion

The study of the effect of prenatal cocaine exposure and the social environment

on the behavior and development of children requires a research design that incorporates

the methodological strengths found in experimental research and a theoretical framework

that allows one to examine numerous levels of influence on an outcome. This is why I

advocate the use of a conceptual model that integrates social, behavioral, and biomedical

methodologies and theoretical perspectives. Despite the contributions biomedical,

behavior, and social science researchers have made to the literature on the impact of








prenatal cocaine exposure on the developing fetus and infant, there is still a knowledge

gap to fill.

Since the mid-1980s, numerous biomedical studies' (Azuma and Chasnoff 1993;

Bellini, Massocco, and Serra 2000; Chiriboga, Brust, Bateman, and Hauser 1999; Fantel

and Macphail 1982; Garavan, Morgan, Mactutus, Levitsky, Booze, and Strupp 2000;

Roland and Volpe 1989; Volpe 1992; Vorhees 1995; Zuckerman 1985), behavioral

studies (Alessandri, Sullivan, Bendersky, and Lewis 1995; Eyler and Behnke 1999;

Leech, Richardson, Goldschmidt, and Day 1999; Lester, Corwin, Sepkoski, Seifer,

Peucker, McLaughlin, and Golub 1991; Richardson, Hamel, Goldschmidt, and Day 1996)

and development studies (Arendt, Singer, Angelopoulos, Bass-Busdiecker, and Mascia

1998; Coles and Platzman 1993) have reported on the negative effects of cocaine on fetal,

neurological, physiological, cognitive, and behavioral outcomes. However, with the

exception of a few recent studies and reviews (Hubbard 1998; Inciardi, Surratt, and Saum

1997; Mayes 1992; Smith 1992), very little sociological or multidisciplinary research has

been done on children who have been prenatally exposed to cocaine. Incorporating a

multidisciplinary perspective will help further our empirical understanding of the social,

structural, environmental, and psychosocial factors that shape children's development

and that influences their behavior.









SThe included references exemplify studies published and cited repeatedly over the past
twenty years. In no way is it an exhaustive list of all the research produced on prenatal
cocaine exposure.













CHAPTER 2
MULTIDIMENSIONAL MODEL



The plethora of contradictory and inconclusive findings in the biomedical and

behavioral science literature on the effects of prenatal cocaine exposure is demonstrative

of how the existing knowledge gaps are best filled by a multidimensional approach. In

this study, I apply a multidimensional approach that integrates biomedical, behavioral

and sociological principles. This theoretical process is consistent with the vision of the

new social and behavioral health science research taking place at the national level

(Singer & Ryff 2001). The multidimensional model that I have conceptualized is heavily

influenced by the research agenda established by the Office of Behavioral and Social

Science Research (OBSSR) at the National Institutes of Health (NIH). The OBSSR

agenda priorities the study of "biological, behavioral, psychological, and social

precursors to disease" and "biological, behavioral, and psychosocial factors that

contribute to resilience, disease, resistance, and wellness" (Singer and Ryff 2001: 3).

In this chapter, I outline the multidimensional model and how it will be applied to

the study of the behavior and development of prenatally cocaine exposed children. I

begin by providing a brief background of the sociological perspective, in order to provide

a foundation for the multidimensional model that bridges the gap between the social,

behavioral, and biomedical disciplines. Then I proceed by defining behavior and








development and the rationale for studying pre-school aged children. The chapter ends

with a restatement of the research questions and general hypotheses.


Sociological Perspective

From a sociological perspective, the study of behavior and development typically

involves the study of human interaction and the relationships that evolve from that

interaction. These relationships are manifested in the form of social roles, norms and

groups that dictate human behavior and can influence human development. In sociology,

behavior is more or less viewed in terms of social action or performance and

development is seen within the context of the life course. Micro-sociological theories,

such as symbolic interaction, focus on the factors that drive social performance or action

(Berger & Luckmann 1966), while macro-sociological theories, like the life course

perspective, focus on how larger social forces or period events impact society over time

(Elder 1995).

This study incorporates both micro and macro level concepts. These concepts

facilitate our understanding of how micro level constructs (such as maternal drug use

and/or parent-child interaction) interact with macro constructs (like social and economic

demographics) to predict outcomes for children whose lives have been effected by the

physiological and social consequences of prenatal cocaine exposure. The use of multiple

dimensions to create a more complex understanding of social life and social structures is

the foundation of sociology as a discipline (Parsons 1951). Sociology is not simply the

study of society, but the study of relationships between multiple dimensions of society,

including biomedical, behavioral, and social factors.








Talcott Parsons (1982), a leader in the field argues that sociology should be driven

by analytical theories that:

help the researcher determine which factors are important and which ones are not
important to the research;
provide a basis for organization of the factors included in the analysis;
help uncover the existing knowledge gaps that need to be filled; and
allow for the integration of theories from related disciplines.

The multidimensional approach that drives this research attempts to do these four

things by examining the existing research and integrating theory. The review of the

literature on prenatal cocaine exposure helped me to identify what factors should be

included in the multidimensional model and how they should be organized. Plus, the

literature review allowed me to uncover the existing knowledge gaps, in particular the

lack of an integrative approach and the need for more research on pre-school aged

children's behavior and development. Therefore, my research integrates micro and

macro level constructs from a variety of disciplines as a means of constructing a

multidimensional model of the behavior and development of prenatally cocaine exposed

children.


Components of the Multidimensional Model

I have constructed a multidimensional model of behavior and development that

integrates the biomedical, behavioral, and social sciences. As exhibited in the previous

chapter, the research conducted by biomedical, behavioral and some sociological

researchers present contradictory results on the effects of prenatal cocaine exposure on

the developing child. There is some evidence that prenatal cocaine exposure is

significantly correlated with fetal brain disruption (Bellini 2000), low birth weight

(Cherukuri et al. 1988; Chouteau et al. 1988; Ryan et al. 1987; Zuckerman et al.1989)








and premature births (Cherukuri et al. 1988). Several behavioral research studies have

revealed that prenatal cocaine exposure is related to poor motor development (Arendt,

Angelopoulos, Salvartor, and Singer 1999), cognitive development (Alessandri et

al.1998; Arendt et al 1998), and language development (Angelilli, Fischer, Delaney-

Black, Rubinstein, Ager, and Sokol 1994; Johnson, Seikel, Madison, Foose, & Rinard

1997), yet the sociological literature (or lack there of) reminds us that there are still a lot

of unanswered questions about the relationship between the prenatal and postnatal

influences on developmental and behavioral outcomes for these "at-risk" children.

I intend to employ a sociological perspective to examine not only how prenatal

cocaine exposure relates to two particular outcomes, but I will examine how prenatal

cocaine exposure impacts these outcomes when it is included in a conceptual model that

includes neonatal and postnatal factors. The conceptual model of behavior and

development presupposes that prenatal and neonatal physiological factors and postnatal

behavioral and social environmental factors each play a role in determining the behavior

and development of prenatally cocaine exposed children as they age. This

multidimensional approach provides a consummate strategy for studying such a complex

problem.

The multidimensional model I have created combines the tenets of sociology,

ecology, psychology, and biomedicine by defining social action, the sociological concept,

as the biological mother's use of cocaine during pregnancy and by using biomedical,

behavioral, and sociological constructs to describe the biomedical and social ecological

factors included in the model. The prenatal and neonatal dimensions represent the

biomedical factors and the social ecological factors represent the postnatal dimension.








The neonatal and postnatal factors help mediate the effect of the prenatal dimension on

behavior and development, two psychologically oriented constructs.

Prenatal Dimension

First, the negative effect of cocaine could be a result of the physiological impact

cocaine has on the developing fetus or it could be due to the fact that many children born

to cocaine using mothers are at a greater risk of reaping the negative social consequences

of their mother's cocaine habit. These consequences include biomedical and sociological

risk such as, lack of proper health care and nutrition, exposure to numerous drugs during

and post pregnancy, the risk for physical and emotional abuse, poverty, high residential

mobility, and poor or limited parent child interaction due to incarceration, limited

sobriety, or the termination of parental rights because of the mother's drug use.

Therefore, regardless of whether or not prenatal cocaine exposure has a physiological or

sociological effect on children, it is a significant predictor in the multidimensional model

of behavior and development.

Neonatal Dimension

The second factor included in the model is the measure of the neonatal

environment. The neonatal environment is characterized by the child's health status at

birth. According to the biomedical model, the complications encountered during

pregnancy and during labor and delivery are indicative of a child's risk for future health

related problems (Hobel, Youkeles, & Forsythe 1979); therefore, it is an important

variable in the multidimensional model of adaptive behavior and development. Simply

evaluating the relationship between the neonatal environment (prenatal cocaine exposure)

and behavior and development, negates the role that neonatal risk factors and postnatal

social factors play in determining a child's outcomes.








Postnatal Dimension

The social environment, which characterizes the postnatal environment is the

third dimension in the multidimensional model. The potential social influences on

behavior and development are numerous. Bronfenbrenner's social ecological model

serves to organize the most salient factors, including the caregiver, home, family, and

social structural variables. Each of these variables represent what Bronfenbrenner calls

the micro-, meso-, exo-, and macrosystems in his ecological model of human

development (Bronfenbrenner 1979).


Theoretical Framework

Social Ecological Theory

Bronfenbrenner's (1979) ecological model of human development provides a

foundation for describing the relationship between multiple factors and their influence on

an outcome. Using the principles of systems theory, Bronfenbrenner's (1979) ecological

approach describes how a series of environmental systems create dimensions of influence

on human development and ultimately behavior. In his book, The Ecology of Human

Development, Bronfenbrenner defines the ecology of human development as such:

The ecology of human development involves the scientific study of the
progressive, mutual accommodation between an active, growing human
being and the changing properties of the immediate settings in which the
developing person lives ... (Bronfenbrenner 1979: 21).

In other words, ecology is the study of humans and their relationship to the environment

in which they live. The definition offered by Bronfenbrenner reminds social scientists

that the relationship that is being observed is not static, but dynamic and it is subject to

change as human and environmental properties change.








Bronfenbrenner goes on to state: "... this process is affected by relations

between settings, and by the larger contexts in which the settings are embedded"

(Bronfenbrenner 1979: 21). He describes the observed relationship between humans and

their environment as a system that is embedded within other systems similar to a set of

Russian dolls. The smallest doll represents the most immediate setting or the

microsystem. The microsystem is encompassed by the mesosystem proceeded by the

exosystem. Last, the entire system is influenced individually and collectively by the

macrosystem reflecting a higher level of influence, such as culture or subculture

(Bronfenbrenner 1979).

The microsystem has been defined as the psychosocial status of the child's

primary caregiver. According to child development theorists, there is a direct

relationship between parental attitudes and behaviors and children's outcomes (Cochran

and Brassard 1979). Additionally, Zuckerman and his associates (1990) found a

significant relationship between early childhood problems and maternal depression even

when controlling for cocaine exposure status. Bronfenbrenner (1979) refers to this

relationship between parents and their children as a dyadic relationship that functions in a

reciprocal fashion.

For instance, a mother who experiences depression may be less responsive to her

child's needs resulting in poor behavioral outcomes for the child. Reciprocally, a child

who presents numerous behavioral and developmental problems may contribute to

increased feelings of helplessness and depression for the mother. Researchers have found

that maternal depression is related to poor child rearing practices (Coletta 1983) and that

poor parental efficacy is corrTelated with poor behavioral outcomes for children (Johnston








and Mash 1989). Hence, the psychosocial status of the primary caregiver is included in

the postnatal factor of the multidimensional model as a means of assessing its

relationship to behavior and development.

The mesosystem is defined as the nature and quality of the child's home

environment. Like the microsystem, the mesosystem is presumed to influence behavior

and development. Studies on the home environment have shown that children who live

in stimulating and loving environments can overcome numerous obstacles, such as

developmental delays caused by poor neonatal health outcomes (Mayes 1996), foster care

placement (Horwitz, Simms, and Farrington 1994), parental divorce, and poverty (Miller

& Davis 1997). According to Bronfenbrenner (1979), children's developmental potential

is enhanced when their physical and social environment enables them to participate in a

variety of activities.

The family's social support mechanism reflects the exosystem and the families'

sociodemographic characteristics comprise the macrosystem in the postnatal

environmental factor. Bronfenbrenner's ecological model of human development clearly

states that the relationship between the family and its support mechanisms and the

family's access to support mechanisms is a critical part of human development; hence

these are essential parts of the multidimensional model that has been constructed to

evaluate the behavior and development of the children in this study. According to

Bronfenbrenner (1979: 7):

whether parents can perform effectively in their child-rearing roles
within the family depends on role demands, stresses, and supports
emanating from other settings. As we shall see, parents' evaluations of
their own capacity to function, as well as their view of their child, are
related to such external factors such as flexibility of job schedules,
adequacy of child care arrangements, the presence of friends and








neighbors who can help in large and small emergencies, the quality of
health and social services, and neighborhood safety.

The factors that characterize the exosystem and the macrosystem are related to

parental behavior, which in turn is related to childhood behavioral and developmental

outcomes. Parents who have a greater sense of family and social support will experience

less stress, thus enhancing their relationship with their children. The improved

relationship between parent and child leads to improved behavioral development

(Cochran and Brassard 1979; Dunst and Trivette 1986; Johnston and Mash 1989).

Additionally, the social and economic factors that characterize the macrosystem

serve as proxies for the availability of flexible job schedules, the quality of care, and the

relative safety of the neighborhood. These factors in the postnatal environment also have

an effect on parental behavior; thus having an impact on childhood outcomes. Families

who have limited economic resources are more likely to have less flexible jobs, poorer

quality health care, and to live in economically disadvantaged neighborhoods plagued by

crime and other safety hazards.

The prenatal, neonatal and postnatal dimensions of the multidimensional model

were determined by integrating biomedical, behavioral, and sociological theories, as

encouraged by Talcott Parson's (1982), who called for sociologists to integrate theories

from a multitude of disciplines. He also argued that sociology should be driven by

analytical theories that provide a basis for how the factors in a model are organized. I

have organized the multidimensional model using the compensatory model of resilience

introduced by Zimmerman and Arunkumar (1994). The prenatal environment represents

the biomedical risk to early childhood outcomes while the neonatal and postnatal

biomedical, behavioral, and sociological factors compensate for the impact of the









prenatal risk. It is presumed that the positive attributes found in the neonatal and

postnatal environments will contribute to the resiliency of "at-risk" children.

Resiliency Theory

Resiliency theory is useful in understanding individual development and behavior

(O'Connor and Rutter 1996; Rutter 1989). According to resiliency theorists, there are

two central theorems that illustrate the concept of resilience, the compensatory and the

protective models (see Figure 3.2). Traditionally, resiliency is defined as the ability to

recover after experiencing some type of distress or misfortune. This distress is referred to

as a risk factor in resiliency models. Resiliency researchers strive to identify the factors

that either compensate for the distress (compensatory model) or that protect individuals

from the distress (protective model) resulting in positive outcomes (Davis 1999;

Zimmerman and Arunkumar 1994).

Compensatory Model Protective Model


RISK -- OUTCOME OUTCOME
FACTOR OUTCOME PI. KOU
FA( I (1<


CO\1I'I \SATORY
FACTOR
PROTECT IVE
FACTOR



Source: Zimmerman and Arunkumar (1994)


Figure 3.1: Compensatory and Protective Models of Resilience



According to the compensatory model, the compensatory factor mediates the effects of

the risk factor by providing additional support or strength. Both factors directly influence








the outcome. This differs from the protective model, in that the risk and the protective

factors interact with one another to influence the outcome. In the protective model, the

protective factor intervenes by reducing or eliminating the risk factor.

In this study, I hypothesize that the observed outcomes, behavior and

development, are directly influenced by two compensatory factors, neonatal risk and the

postnatal social ecological system. In the multidimensional model, I propose that

prenatal cocaine exposure is directly related to the behavior and development of pre-

school aged children, but I also propose that it is not the only factor that has an impact on

the outcome.

As shown in Figure 3.2, the compensatory factors play a significant role as well.

I contend that the compensatory factors are directly related to behavior and development.

This theory is driven by my belief that the neonatal and postnatal environments do not

intervene as the protective model presumes, but these biomedical and sociological

factors, actually compensate for the prenatal risk in an additive fashion. Prenatal cocaine

exposure, neonatal health status, and the postnatal social ecological factors each play a

part in determining the behavior and development of pre-school aged children.

Therefore, prenatally cocaine exposed children may be able to overcome the risk

associated with prenatal cocaine exposure when the neonatal risk is reduced and the

postnatal environment is optimized.



















Preinatal Risk
(Cocaine Exposure


I Red = Risk Factor; Green = Compensatory Factor; Blue= Outcome


Figure 3. 2: Multidimensional Model of Behavior and Development




Definitions Of Behavior And Development

Behavior and development can refer to many different constructs and both terms

have been defined and measured in different ways. In this research, I define and measure

behavior and development separately. The term behavior actually refers to the children's

adaptive behavior or their ability to demonstrate the skills and perform the tasks that are

expected of them as they reach maturity. Behavior reflects the acquisition of the

necessary skills for taking care of oneself and getting along with others (Vineland 1984).

The study of adaptive behavior among prenatally cocaine exposed children is important

because it serves as a proxy for the child's ability to survive in a complex world that

requires independence, as well as social cooperation or the ability to conform to social

norms.


Neonatal
Health Risk
Dimension



Behavior and
SDevelopment





Postnatal
"Social n-toil"nsic;l"
Dimension








Development refers to the achievement of particular growth parameters in

relation to chronological age. The typical measures of development include abstract

reasoning, memory, learning, and problem-solving abilities and fine and gross motor skill

development. These skills play an important role in a child's ability to orient towards the

environment and they influence the quality of the child's interaction with the

environment (Bayley 1969). In this study, development and behavior are measured

independently, but they are presumed to follow the same pathways. Some developmental

scientists would argue that the behavior and development are inextricably intertwined

(see Bergman, Cairns, Nilsson, & Nystedt 2000; Bijou and Ribes 1996; Immelmann,

Barlow, Pterinovich, & Main 1981; Pujol, Vendrell, Junque, Marti-Vilalta, & Capdevila

1993 for discussions on behavior development); hence, the application of one model for

both constructs.

Reason for Studying Pre-Schoolers

Half of the children included in this study of adaptive behavior and development

have been prenatally exposed to cocaine and the other half have not been exposed to

cocaine prenatally, but both groups can be defined as "at risk" of poor behavioral and

developmental outcomes. The majority of the sample is comprised of Black children

born of relatively inexperienced and poorly educated mothers. By age three, many of the

children are living with primary caregivers who are at-risk of depression and who have

household incomes below the poverty line.

Thirty-six months is a prime age to observe the behavior and development of

these "at-risk" children, because as children reach the age of three, they are expected to

perform tasks and acquire skills that are not easily measured or observed during earlier

ages. According to the American Academy of Pediatrics, three-year old children are








increasingly social and inquisitive, thus enhancing their independence, which is marked

by their attempts at communication and mobility (Cowley 2000; Kantrowitz 2000;

Raymond 2000). The acquisition of social, communication, daily living, mental and

motor skills are important to the child's overall growth and development; however for

many poor, Black children their growth and development is thwarted by their

impoverished social and economic status (Brooks-Gunn & Duncan, 1997; McLeod &

Nonnemaker, 2000; McLoyd, 1990).

Studying behavior and development at age three makes methodological sense

because, most children have not entered a formal educational institution; therefore, the

influences on their behavior and development are considerably less than school age

children. Even though three-years is a relatively short period of time given the typical

age of mortality in the United States, three-year olds do have a life course trajectory.

According to the multidimensional model, their prenatal, neonatal, and postnatal

environments reflect a trajectory that impacts their behavior and their development as

they mature. Each environmental dimension represents a particular level of influence in

the multidimensional model. The neonatal and postnatal environments represent levels of

influence that compensate for the risk associated with a harmful prenatal event,

specifically prenatal drug exposure.


Research Questions

The conceptual model outlined in this chapter will be used to answer the

following questions:

How do prenatally and non-prenatally cocaine exposed children differ on
measures of behavior and development at age three?








To what extent do prenatal, neonatal and postnatal factors, including
prenatal cocaine exposure, neonatal health and the postnatal social
ecological factors, enhance or hinder pre-school age children's behavior
and development?

I expect that when controlling for the neonatal and postnatal factors, the

impact of prenatal cocaine exposure on the behavior and development of pre-

school aged children will be reduced. Previous researchers have found significant

differences between prenatally and non-prenatally cocaine exposed infants,

whereas the exposed infants are at a greater risk of experiencing poor outcomes.

But does the effect of prenatal cocaine exposure persist beyond infancy and into

early childhood? And if so, can these effects be compensated for when the child

has minimal health problems and resides within a positive social ecological

setting? The anticipated answers to these questions are outlined in the next

chapter, which also includes a detailed description of the research methods, the

hypotheses and how they are operationalized.













CHAPTER 3
RESEARCH VARIABLES AND SAMPLE DESCRIPTION



Designing a quality study involves constructing a research protocol that contains

valid and reliable measures of the variables outlined in the theoretical model. This helps

to minimize the bias that results from poorly designed studies. After creating the

conceptual framework outlined in the previous chapter, I was fortunate enough to find a

source of data that contained the variables necessary to address the questions posed in the

previous chapter. In addition to the data source being comprehensive, the procedures

used to collect the data are empirically reliable and valid to the extent that researchers are

able to depend on the honesty and integrity of human subjects. In this chapter, I describe

the data collection techniques, the longitudinal sample, and the operationalization of the

prenatal, neonatal, and postnatal factors in the multidimensional model.


Data Collection

The data used in this study comes from a cross-section of the data collected for a

longitudinal study on cocaine use among pregnant women conducted in the Department

of Pediatrics at a university hospital in the southeastern region of the United States. From

1991 to 2001, data were collected from study participants during their pregnancy, at the

birth of the target child, and during scheduled follow-ups with the target child and his or

her respective primary caregiver. The data used in this study come from the follow up

interviews and assessments conducted when the target child reached 3 years of age.








This study uses a quasi-experimental design to assess the adaptive behavior and

development of children who have been prenatally exposed to cocaine by comparing

them to a group of matched controls. The recruitment procedures were based on a

matched group design in order to minimize the effects of possible confounding variables,

such as socioeconomic status, race, parity, and level of prenatal risk (Behnke, Eyler,

Woods, Wobie, and Conlon 1997).

Recruitment took place between 1991 and 1993 at the recruitment hospital or at

one of its surrounding prenatal clinics. The researchers approached a total of 2,526

potential participants during their first or subsequent prenatal visit or at delivery in order

to include a full range of pregnant cocaine users. Only women less than eighteen years of

age, who did not speak English, who used an illicit drug other than cocaine and marijuana

and/or confounding prescription medications, and those who were diagnosed with a non-

drug related chronic illness known to effect pregnancy outcomes were excluded (e.g.

diabetes, sickle cell). Of the 2,526 women recruited, 85% (n-2147) consented to

participate in the study upon first approach. Upon consent, each woman was given a

urine drug screen and a drug history interview.

A power analysis was conducted to determine the necessary sample size in order

to maintain statistical power of p = 0.05 and to detect an effect size within one-third of a

standard deviation on developmental outcomes (i.e. the Bayley Scales for Infant

Development). Based on the results of the power analysis it was determined that an

enrollment of 150 cocaine users and 150 matched controls would be sufficient assuming

an attrition rate of 33% over three years (Eyler, Behnke, Conlon, Woods, and Wobie

1998). Hence, the first 179 women with positive drug screens and/or admitted cocaine








use at the time of delivery were asked to continue their participation in the study. Of

these women 154 consented, fit the inclusion criteria and were positively identified as

prenatal cocaine users based on their positive urine drug screen given on the day of

enrollment and at delivery. One hundred and fifty-four women with negative drug

screens and no admission of drug use were enrolled in the control group using the

matching criteria.

Urine specimens for the mother and infant were tested using fluorescence

polarization immunoassay. Positive screens were confirmed through the use of gas

chromatography-mass spectroscopy. This process involves the use of a magnetic device

and computer technology to isolate the ionic components of the substance being tested.

The use of two drug screens and detailed drug histories were employed to minimize the

chances of misclassification of drug users as non-drug users (Behnke, Eyler, Woods,

Wobie, and Conlon 1997). These methods of classification are highly regarded by

researchers in the field (Arendt, Singer, Minnes, and Salvator 1999). The University's

Institutional Review Board approved all methods and procedures and great care was

taken to assure that each participant understood what she was agreeing to when signing

the informed consent and to protect the confidentiality of all participants.


Description of Longitudinal Sample and Sample Retention

Description Of Birth Mothers

At birth the total sample included 154 cocaine users and 154 non-cocaine users.

Each woman was interviewed prior to the birth of the target child and/or immediately

after the child was born. When possible each woman was first interviewed during her

initial entry into the health care system and during each subsequent trimester to facilitate








her recent recollection of drug use and other life events. Forty-one percent began their

interviews during the first trimester, 34% during the second trimester and 25% of the

women did not receive their first interview until delivery as a result of when the subject

entered the health care system. The women enrolled at delivery received limited or no

prenatal care. Women who were trained to establish rapport conducted the prenatal

interviews. The interviewers were instructed to appear non-judgmental, supportive, and

encouraging. This enhanced the reliability of the information collected.

The group of cocaine users is not significantly different from the non-cocaine

users on any of the matching criteria. The two groups were matched on race,

socioeconomic status, parity, and perinatal risks. On average, the cocaine users were

older than the non-users by 3 to 4 years, yet both groups fell within the same age range of

18-43 with only two women (one in each group) over the age of 40 years. The majority

of the cocaine users were Black (n=125). The majority of the cocaine users were

classified as unskilled laborers as determined by the Hollingshead Index (Hollingshead

1975) and the majority had received 12 years or less of formal education. The cocaine

users and non-cocaine users reported some alcohol, marijuana, and tobacco use with the

cocaine users reporting significantly higher usage rates on all substances. Forty-two

percent of the sample reported never being married. Additionally, 134 of the cocaine

users reported having previous births; hence they were matched with multiparous non-

cocaine users. Table 4.1 provides a description of the biological mothers. Non-

parametric statistical procedures were used to calculate the mean differences between the

two groups. Each table reports the mean differences, the mean rank, and the Mann-

Whitney p-value.








Table 4. 1: Descriptive Statistics of Biological Mother and Child at Birth by Cocaine
Exposure Status, Means +/- Standard Deviations (Mean Rank)
Non- Cocaine Total Mann-
Cocaine Exposed Sample Whitney
Exposed Test

n=154 N=308 p-values
________________n=154_____

Age in Years 23.8+/-5.5 27.6+/-4.8 25.8 <0.001

Years of Education 11.4+/-1.1 11.2+/-1.5 11.3 +/-1.29 0.440
Completed (157.72) (150.25)______

Race= Black 0.81 0.81 0.81+/-0.39 1.000
(154.50) (154.50)_______

Symptoms of 23.5 29.7 26.6 <0.001
Depression (CES-D)______
Socioeconomic
Status
Professional (1) 0 0 0
Mid-level Prof (2) 1(0.7%) 1(0.7%) 2 (0.6%) 0.994a
Skilled Worker (3) 3(2.1%) 3(2.1%) 6(1.9%)
Semi-skilled (4) 29(20.1%) 31(21.5%) 60(19.5%)
Unskilled (5) 111(77.1%) 109 (75.7%) 220(71.4%)

Missing values 10 10 20 (6.5%)
a Chi-Square Test


Characteristics Of Children At Birth

Within the longitudinal sample of 308, there were several differences between the

prenatally cocaine exposed and non-exposed children's birth outcomes. There were three

fetal deaths among the women who used cocaine in utero and one death in the control

group; however, the difference was not statistically significant (p = 0.62). As shown in

Table 4.2, the women who used cocaine in utero were twice as likely to have a premature

birth.








Table 4. 2: Longitudinal Study Sample Birth Outcomes by Cocaine Exposure Status,
Means +/- Standard Deviations (Mean Rank)
Non-Cocaine Cocaine Total Mann-
Exposed Exposed Sample Whitney
Test
p-values

n=154 n=154 N=308
Infant's
Gender = 0.45 0.55 0.47+/-0.50 0.077
Female (146.29) (161.76)
n=154 n=154 N=308
Premature
Infant = Yes .09 .19 .14+/-0.35 0.014
(114.83) (120.26)
n=153 N=149 n=302
Infant's Birth 0.003
Weight 3179.4+/-699.6 2984.9+/-668.2 3083.4+/-
(166.1) (136.5) 690.0_____
Hobel Risk n=151 N=148 n=299
Score
Prenatal 43.0 +/-19.3 54.9 +/-20.6 48.9+/-20.8 <0.001
(124.47) (176.04)

Labor & 18.9+/17.2 18.1 +/14.1 18.6 +/-15.7 0.780
Delivery (148.63) (151.40)

Neonatal 16.6 +/-36.2 21.6 +/-61.5 19.1 +/-50.3 0.135
(142.64) (157.51)________________
'The missing values are attributed to fetal death and missed assessments for
children born outside of the hospital.


Nineteen percent (n=29) of the prenatally cocaine exposed neonates were born

prematurely compared to 9% (n=14) of the non-prenatally cocaine exposed neonates.

This difference is statistically significant at p=0.014.

Also, there is a statistically significant difference between the prenatally cocaine

exposed children and non-prenatally cocaine exposed children's mean birth weight (p =

0.003). On average the birth weight of the PCE children is lower (mean = 2984 grams;

standard deviation = 668) than the non-PCE (mean = 3179 grams; standard deviation =








670). Newborns who weigh less than 2500 grams at birth are considered low birth

weight, placing them at risk for physical and developmental complications and even

death. Twenty-two percent (n=33) of the PCE children weighed less than 2500 grams at

birth compared to 11.8% (n=18) of the non-PCE children. As shown in Figure 4.1, the

percentage of low birth weight babies in this sample is considerably higher than the 1994

national average of 7.3%, 13.0% for Blacks, and 6.2% for whites (Ventura 1994).





100.000o '

80.000oo.

60.00% [i Non-PCE
,.- ilPCE
40.00% .- IEPCEI

20.0000.

0.00%0
Above 2500 Below 2500


Figure 4. 1: Percent of Low Birth Weight Children by Prenatal Cocaine Exposure Status,
n=302



Even though there were differences in the average birth weight of the PCE

children and non-PCE children, their Hobel Perinatal Risk Assessment scores did not

differ considerably. Medical professionals assessed the risk of complications prenatally,

during labor and delivery, and postnatally using the Hobel scoring system (Hobel 1973;

Hobel 1979). The cocaine-exposed infants (mean=54.5; standard dev. =20.1) had

significantly higher prenatal risk scores than the controls (mean =43.0; standard dev.








19.3), but there was no statistically significant difference on the labor and delivery (p =

0.780) or neonatal scores (p = 0.135).

Retention And Attrition At Three Year Follow Up

During the three-year follow up, the researchers collected data from each primary

caregiver of the child during a home interview and from each child during a clinic

appointment. The home interviews that were conducted with each primary caregiver

included measures of the caregivers' psychosocial status (CES-D), the home environment

(HOME), the families' social support system (FSS), and the children's adaptive behavior

(Vineland). A trained interviewer conducted the home interviews. Also, a blinded

clinician on or around the child's third birthday completed the child's developmental

assessment (Bayley).

Of the initial 308 subjects, 80.2% (n=247) received both a home and a clinical

assessment at age three. Ten of the children had died by age three and 14 families were

lost to follow up or declined. Nine subjects completed either the home interview or the

clinical assessment, but not both, and 39 missed some part of either the home interview

or the clinic. As shown in Table 4.3, there is no significant difference between cocaine-

exposed and non-cocaine-exposed groups in terms of their participation or lack of

participation in the three year follow up (chi-square = 3.869 [4]; p=0.424). Additional

analysis was done to compare the birth outcomes of the children who remained in the

study to those who did not remain. There were no statistically significant differences

between the initial sample's mean birth weight, Hobel complications assessment scores,

gender, or prematurity status. Therefore, the 234 subjects with complete data are

included in each bivariate and multivariate analysis with minimal risk of attrition bias.








Table 4. 3: Subject Retention and Attrition at 3-Year Follow Up
Participation in Study at Control Target Total
3 year Follow Up n= 144 n= 138 n= 282
Completed Both 123 124 247
39.9% 40.3% 80.2%
Missed Clinic or Interview 6 3 9
1.9% 1.0% 2.9%
Incomplete Clinic or 15 11 26
Interview 4.9% 3.6% 8.4%
Reasons for Exclusion Control Target Total
from 3 year Follow ULip n=10 n=16 n=26
Child Died Prior to 3 year 5 5 10
Follow-up 1.6% 1.6% 3.2%
Unable to Locate for 5 11 16
Follow-up 1.6% 3.6% v5.2%
Total 154 154 308
__________50.0% 50.0% 100.0%
Pearson Chi-Square=3.869 (df=4) alpha=0.424


Operationalization Of Variables and Hypotheses

Prenatal Risk: Prenatal Cocaine Exposure (PCE)

Prenatal cocaine exposure was determined by multiple measures including

maternal drug histories and urine assays taken from the mother during pregnancy and

from the child at birth. Positive urine screens were confirmed using the gas

chromatography -mass spectroscopy process. Target subjects included any woman who

admitted to using cocaine and all the women and children who had positive confirmed

urine screens for cocaine. Those who denied use and had negative cocaine screens

comprised the control group.

As in many of the studies of prenatal cocaine exposure, polydrug use exists

among our sample of cocaine users. Both the cocaine users and some of the non-cocaine

users reported the use of alcohol, marijuana, and tobacco. Subjects who reported the use

of any illicit substance other than cocaine and marijuana were excluded from the study.








The amount of cocaine used by the cocaine users was determined from oral reports taken

by the interviewer. Each cocaine user was asked how often they used cocaine, how much

cocaine they used, or approximately how much money they spent on cocaine. The

women in this study were primarily crack cocaine smokers. Their cocaine use ranged

from smoking during as little as 2% of the time they were pregnant to as high as 100% of

the time. On average, the cocaine users smoked crack during approximately 48% of their

pregnancies (standard deviation = 0.28). Figure 4.2 is a histogram of the percent of

weeks cocaine was used during the pregnancy for the cocaine users in the sample. Table

4.4 shows the comparisons between the cocaine and non-cocaine users drug habits

including the cocaine use for the initial longitudinal sample.

The prenatal risk factor describes the children who were and who were not

exposed to cocaine prenatally. All of the children of mothers who used cocaine during

their pregnancy, regardless of the amount or timing of the use, were included in target or

PCE group. This study is not a study of the variable effect of prenatal cocaine exposure,

but a study of the differences between children who have been deemed "at-risk" due to

prenatal cocaine exposure. Typically, the "at-risk" label is used without knowledge of

maternal drug use during pregnancy, let alone the amount of cocaine used. Therefore, it

in this particular analysis it is more appropriate to study the children who are "at-risk" of

being labeled as "at-risk" than it is to study only the children with the greatest exposure

to cocaine.










Histogram

For TARGET= target


0.00 .13 .25 .38 .50 .63 .75 .88 1.00
.06 .19 .31 .44 .56 .69 .81 .94


Percent of weeks Cocaine Used During Pregnancy


Figure 4. 2: Percent of Weeks Cocaine Used During Pregnancy, n=115


Table 4. 4: Biological Mothers Drug Use by Cocaine Exposure Status, n=308
Non- Cocaine Total Chi-Square
Exposed Exposed Sample Test

n=154 N=308 p-values
_________________n=154_______________
Percent of Weeks
Cocaine Used 0.47+/-.30 -- --
During Pregnancy__________

Marijuana Use 7.1% 44.2% 25.6% .000
(11) (68) (79)

Tobacco Use 24.0% 79.9% 51.9% .000
(37) (123) (160)

Alcohol Use 30.5%' 76.6% 53.5% .000
~____~__(47) (118) (165)_____


20






10'



0-.
CT"
0
fL 0.


Std. Dev = .28
Mean = .48
N= 115.00








Outcome Measures: Behavior and Development

The multidimensional approach outlined in the previous chapter will be used to

assess the differences in the behavior and development of prenatally and non-prenatally

cocaine exposed children. The score on the Vineland Adaptive Behavior Scales

(Sparrow, Balla, & Cicchetti 1984) will be used as a measure of behavior. The score on

the Bayley Scales of Infant Development will be used to measure the children's

development.

Behavior. The Vineland scale was designed to measure the personal and social

adaptability of handicapped and non-handicapped individuals from birth through

adulthood (Holden 1984). During a visit to each child's home, the instrument was

administered to each child's primary caregiver in the form of a semi-structured interview.

The Vineland was scored and interpreted by a trained test administrator upon its

completion.

The Vineland survey consists of 297 items that cover four behavioral domains:

(1) communication (receptive and expressive language); (2) daily living skills (self-care

activities); (3) socialization (interpersonal relations and play activities); and (4) motor

skills (gross and fine motor coordination). The standard scores range from 0 to 200 with

higher scores reflecting higher levels of competence (Sparrow, Balla, and Cichetti 1984).

Each primary caregiver reported whether or not her child performed the tasks

satisfactorily and habitually or if the skill was emerging and/or adequately performed. If

there had not been an opportunity to observe the child performing the tasks, the task was

not included in the total score for that domain.

The standard scores for the Vineland were developed using a national

standardized sample of 3,000 subjects drawn from a pool of 21,876 potential participants








representing the 1980 US Census figures by race, sex, community size, geographic

region, and parental education (Holden 1984). The developers of the Vineland report

test-retest correlations ranging from 0.98 to 0.99 and interrater reliability correlations

ranging from 0.96 to 0.99 (Sparrow, Balla, and Cichetti 1984). The construct validity

was reported as satisfactory based on an increase in raw scores for each age group,

significant factor loading, and positive correlations with other intelligence scales, such as

the Peabody Picture Vocabulary Test (Sparrow et al. 1984).

Development. The Bayley Scales of Infant Development (Bayley 1969) were

designed for use in both clinical and research practices by Nancy Bayley in 1969. It

provides an assessment of each child's developmental status in comparison to national

norms. The scales were standardized on a stratified sample of non-institutionalized

children selected from the 1960 United States Census. Almost 40 years after its

inception, the Bayley Scales are still widely used among researchers. The primary

strengths of the instrument are its norms that controlled for sex, race, urban versus rural

residence and parental education, as well as its significant correlation with other

intelligence scales such as the Stanford Binet (Whaley 1984).

A study by Werner and Bayley (1966) reports respectable statistical properties for

the Bayley. Split-half reliabilities were reported for the mental and motor scales. The

Spearman correlation coefficients for the mental scale ranged from 0.81 to 0.93 with a

median of 0.88. The motor scale has resulting coefficients of 0.68 to 0.92 with a median

value of 0.84. The test-observer and test-retest reliabilities were also favorable. The

mean percentage of agreement between observers on the mental scale was 89.4 (standard

deviation = 7.1) and 93.4 (standard deviation = 3.2) on the motor scale. The test-retest








reliability score for the mental scale was reported as 0.76; however, the motor score was

not reported in the review (Whaley 1984).

In this study, a clinician completed the Bayley during the children's three-year

follow-up at the hospital or clinic. The clinician used the Bayley to assess each child's

mental and motor skills. Like the Vineland, the standard scores on the Bayley range from

0 to 200 where higher scores reflect a higher level of competence (Bayley 1969). The

mean on the mental and motor index is 100 with a standard deviation of 16. Interpreters

of the Bayley and the Vineland have placed the scores in the following range: Low (less

than 69), moderately low (70-84), adequate (85-115), moderately high (116-130), and

high (131 or above).

Hypothesis 1: Based on my review of the biomedical and behavioral
literature on the effects of prenatal cocaine exposure, I hypothesize that in
a bivariate analysis, the children who have been prenatally exposed to
cocaine will have lower scores on both the Vineland and the Bayley than
children who have not been prenatally exposed to cocaine. However, this
prediction changes when the neonatal and postnatal factors are considered.

Neonatal Risk: Newborn Health Status

Based on biomedical and behavioral research there is a relationship between poor

health at birth and poor developmental outcomes in infancy and in some cases early

childhood. Each child's risk of experiencing physiological problems that require medical

attention was assessed at birth using the Hobel Perinatal Complications Scale (Hobel,

Hyvarinen, Okada & Oh 1973). Medical personnel determined each child's neonatal risk

after reviewing the mother and child's medical records, which included an assessment of

prenatal risk, neonatal risk, and an assessment of the complications incurred during labor

and delivery. The scores on these assessments indicate each child's risk of experiencing

health problems that may require medical attention in the future.








The neonatal risk score ranges from 0 to 35. A score of 10 or higher indicates a

greater risk of having physiological problems during infancy and beyond. In order to test

the reliability of the instrument, Hobel and his associates (1979) tested the scale's ability

to predict infant mortality. They found that the neonatal risk score predicted infant

mortality correctly 82.5% of the time (Hobel, Youkeles & Forsythe 1979). Hence,

children with lower risk scores should experience better health outcomes, thus

compensating for the risk associated with prenatal cocaine exposure.

Hypothesis 2: Using the multidimensional model, I predict that regardless
of cocaine exposure status, fewer neonatal risks will result in higher scores
on the adaptive behavior and developmental scales. Additionally, the
neonatal risk will serve as a significant predictor of adaptive behavior and
development when included in the multidimensional model.

Postnatal Factors: Social Ecological Model

The postnatal environment reflects the social ecological model adapted from

Bronfenbrenner's (1979) work. Using Bronfenbrenner's ecological model of human

development as a guide, I have constructed four systems or dimensions that characterize

the postnatal environment: the micro-, meso-, exo-, and macrosystems. The

microsystem represents the setting closest to the child and it is operationalized using a

measure of the primary caregiver's psychosocial status. The mesosystem characterizes

the relationship between the child and his or her home environment. According to

Bronfenbrenner, the exosystem reflects the setting that does not directly involve the child,

but may have an influence on the child's development. It is operationalized as the

family's social support system. Last, the macrosystem is comprised of a measure of the

child's race and the family's household income.

At age three, 26.5% (n=62) of the children in the total sample do not live with

their biological mother (see Table 4.5). In each case, the individual who had custody








and/or who was primarily responsible for the child's care has been classified as the

child's primary caregiver. The responses from the interview with the primary caregiver

are used to construct the variables in the social ecological model, which represents the

postnatal environment, the second compensatory factor in the multidimensional model.




Table 4. 5: Caregiver's Relationship to Child at Age Three by Cocaine Exposure Status
Non- Cocaine Row Z-Test
Cocaine Exposed Percents Statistic
Exposed
n=119 n=115 n=234 p-value
Biological Mother 66.3% 33.7% 100% -7.844
(114) (58) (172) 0.000
Biological Father 12.5% 87.5% 100% -2.203
(1) (7) (8) 0.028
Adopted Mother 100% 100% -2.294
(5) (5) 0.022
Grandmother 13.3% 86.7% 100% -2.998
______(2) (13) (15) 0.003
Other Relative 100% 100% -3.610
_________________ (12) (12) 0.000
Foster Mother 9.1% 90.9% 100% -4.108
______ (2) (20) (22) 0.000


Microsystem: Primary Caregiver's Psychosocial Status. Each child's primary

caregiver's psychosocial status is determined by her score on the Center for

Epidemiologic Studies Depression Scale (CES-D). The CES-D is a self-report of

depressive sympotomology designed for use in research settings with non-clinical or non-

psychiatric populations. It consists of twenty questions rated on a four-point scale. Each

question indicates how frequently the respondent experienced a particular set of

symptoms during the past week. Factor analysis reported by Radloff (1977) indicates

that the CES-D can be broken into four dimensions: positive affect, negative affect,








somatic concerns, and interpersonal concerns. However, given the high degree of

internal consistency, it is recommended that a total score be used in statistical analysis.

A total score of 16 typically points to the presence of depressive symptomology.

A score below 16 suggest that the individual is probably not at risk for depression. A

score between 16 and 20 means the individual has expressed mild depressive

symptomology. A score ranging from 21 to 30 reflects a risk of experiencing moderate

depression and a score of 31 or above indicates severe risk of receiving a diagnosis of

depression. However, it is duly noted that this is not a clinical assessment, but a

diagnostic tool or screen to measure depressive symptoms in the general population

(Devins and Orme 1984).

The test was created in the early 1970s and tested on randomly selected adults

living in households chosen from a probability sample of households in Kansas City,

Missouri or Washington County, Maryland. Numerous researchers have reviewed the

psychometric properties of the CES-D. Initial reports by Devins and Orme (1984) on the

test-retest reliability appear quite low (two week estimate r = 0.51; four week estimate r =

0.67), but it is presumed that the low scores are due to the fact that the scale measures

symptoms that occurred within the last week. So, as people and their life circumstances

change, the results from one week to the next are expected to change as well.

The internal consistency is a more useful measure of reliability for an instrument

that measures changing states or moods such as the CES-D. Using Cronbach's alpha,

Radloff (1977) reports Pearson correlation coefficients of 0.84, 0.85 and 0.90 meaning

that the items in the instrument all appear to be measuring the same construct.








Weissman and associates (1977) tested the construct validity of the CES-D by

testing individuals undergoing psychiatric treatment. It was reported that the depressed

group in the sample received higher scores on the CES-D than the recovered and non-

depressed groups. Additional studies confirm the construct validity of the CES-D by

comparing the results of the CES-D to those of other depression scales. The CES-D is

reportedly highly correlated with the Symptom Checklist 90, another scale designed to

measure psychological distress and depressive symptomology. The correlation

coefficients ranged from 0.70 to 0.80 (Derogatis, Lipman, and Covi 1973).

Hypothesis 3: Using the multidimensional model, I predict that regardless
of cocaine exposure status, the primary caregivers' score on the CES-D
will be correlated with the adaptive behavior and developmental scores for
their child. As the number of depressive symptoms increases, the scores
on the adaptive behavior and developmental skills scales will decrease.
Additionally, the primary caregiver's depression status will serve as a
significant predictor of behavior and development in the multidimensional
model.

Mesosystem: Home Environment. The mesosystem will be measured by

assessing the physical and parenting environments in the home of the child. The physical

and parenting environment will be operationalized using Caldwell and Bradley's (1979)

Home Observation for Measurement of the Environment (HOME) Inventory. The

HOME was completed during the interview with the primary caregiver. The interviewer

completed the HOME by asking each caregiver questions and by observing the family's

behavior and the nature of the environment. The HOME was designed as a standard way

for researchers to ascertain the nature and quality of the physical and parenting

environment as they relate to children's cognitive development.

The HOME scale was standardized on families in Little Rock, Arkansas and

Syracuse, New York. The sample was not random or stratified, but the researchers did








over sample low income and Black families. The revision of the HOME, which took

place in 1984, was tested on 232 families in Little Rock of whom 30j'; were welfare

recipients, 66% were Black, and the average level of education was 11 years of

schooling. The version of the scale designed for children of this age consists of eight

subscales: learning stimulation, academic stimulation, physical environment, variety in

experience, language stimulation, warmth and acceptance, modeling, and acceptance. The

Kuder-Richardson reliability estimate for the total scale is 0.93 and the test-retest

reliability ranges are as high as 0.70 depending on the subscale (Procidano 1985).

For purposes of this study, the home is divided into two factors: the parenting

environment and the physical environment. The parenting environment consists of the

items on the HOME that measure the primary caregiver's interaction with the child

within the home. The physical environment consists of the items that measure the

homes' physical attributes. Two factor analyses were conducted to determine the

viability of these two domains.

The factor loadings for the parenting dimension ranged from 0.56 to 0.78. The

eigenvalue was 2.56 and it explained 51.25% of the variance. The language stimulation,

warmth and acceptance, academic stimulation, modeling, and acceptance subscales,

which make up the parenting factor, maintained statistically significant correlations of

0.74 or greater with the parenting factor.

The factor loading for the physical dimension ranged from 0.66 to 0.81. The

eigenvalue was 1.77 and it explained 58.98'; of the variance. The physical dimension

consists of the learning stimulation, physical environment, and variety in experience








subscales. Each of these maintained statistically significant correlations with the

parenting dimension ranging from 0.66 to 0.82.

In this study, the scores on the parenting dimension range from 1 to 6 with a mean

of 4.26 (standard deviation = 0.89). The physical dimension ranges from 1 to 6 with a

mean of 5.67 (standard deviation = 1.60). The higher the score, the more positive

attributes found in the home environment.

Hypothesis 4: Using the multidimensional model, I predict that regardless
of cocaine exposure status, the parenting environment and the physical
environments will be correlated with the adaptive behavior and
developmental scores. As the number of positive attributes in each
dimension increases, the scores on the adaptive behavior and
developmental skills scales will increase. Additionally, the parenting and
the physical environments will serve as significant predictors of behavior
and development in the multidimensional model.



Exosystem: Family and Social Support System. The exosystem refers to the

quality of the family's social support system. The families' social support system is

measured using the Family Social Support Scale (FSS). The FSS scale measures the

existence of particular family and community support networks and how helpful they

have been to the family during the last twelve months (Dunst 1985). The primary

caregiver is asked to indicate on a 19-item list how helpful immediate and extended

family members and community and professional agencies have been in the past year.

The family members included in the survey include those of the child and the child's

primary caregiver. Using a likert scale ranging from 0 (not at all helpful) to 4 (extremely

helpful) each supporter is ranked with the exception of those who do not apply. The total

score is computed by adding up each rank. In this study, the scores on the Family








Support Scale ranged from 12 to 76 with a mean of 38.43 (standard deviation = 11.96).

The scale is an assessment of the quality of the support available to the primary caregiver.

Dunst, Trivette, and Jenkins (1988) report promising psychometrics from a study

of 139 parents of pre-school children with developmental disabilities. The alpha

coefficient for the measure of validity was 0.77. The split-half reliability measured by

the Spearman-Brown formula was 0.77 and the Pearson test-retest reliability coefficient

after one month was reported as 0.75.

Hypothesis 5: I predict that regardless of cocaine exposure status, the
quality of family support will be correlated with adaptive behavior and
development. As the number of positive supports available to the family
increases, the scores on the adaptive behavior and developmental skills
scales will increase. Additionally, I hypothesize that the family support
dimension will serve as a significant predictor of behavior and
development in the multidimensional model.

Macrosystem: Social Structural Variables. In most sociological research,

demographic characteristics are added into the analysis with very little thought about the

intricate connection between sociodemographics and outcomes. There is enough

empirical evidence available to support the contention that structural variables such as

race, class, and gender do matter (Allport 1988; Davis 1991; Hacker 1992; West 1993).

Bronfenbrenner describes the macrosystem as the setting that represents the culture or

subculture in which the other settings or systems reside. He predicts that the nature of the

micro-, meso-, and exosystems will vary depending on the characteristics of the

macrosystem. For example, "homes, day care centers, neighborhoods, work settings, and

the relations between them are not the same for well-to-do families as for the poor"

(Bronfenbrenner 1979: 26).

In this study, I have constructed four dummy variables that represent the race and

economic status of each family. The race of the family is determined by the race of the








biological mother. Each biological mother was classified as either Black or non-Black

based on self-reports, upon enrollment into the study. I realize this is a crude estimate of

the family's race, but unfortunately, the official racial classification of the children and

their primary caregivers was unavailable. However, one could argue that in most cases a

child's race is determined by his or her biological mother's race (unless the biological

father is Black and the mother is not Black, whereas the child is typically classified as

Black); therefore, using this classification is not completely unreasonable.

The families' household income was collected during the home interview. Each

primary caregiver was asked to classify her families' entire household income into one of

five categories: (1) $0 to $6,000, (2) $6,001 to $12,000, (3) $12,001 to $18,000, (4)

$18,001 to $24, 000, and (5) $24,000 or more. I used these categories to construct a

poverty variable. The poverty variable measures whether or not the family is living at or

below the poverty line based on the 1998 poverty threshold of $16,600 for a family of 4

(NCCP 2000). I took the largest number in each range and divided it by the number of

people living in the household. This number was then measured against the poverty line

and those families with household incomes equal to or below $16,600 were classified as a

"1" meaning below poverty and those who had an income of $16,601 or greater were

classified as a "0" or above poverty.

The poverty variable was then added to the race variable to create four dummy

variables: BlackPoverty (family with a Black child living below poverty), Non-

BlackPoverty (family with a Non-Black child living below poverty), BlackAbove (family

with a Black child living above poverty) and Non-BlackAbove (family with a Non-Black








child living above poverty). The BlackPoverty variable is the referent group in the

multidimensional model.

One of the strengths of the data used in this study is the existence of a matched

sample of cocaine and non-cocaine exposed children. The cocaine using and non-cocaine

using women were matched on race, parity, perinatal risks, and socioeconomic status at

the time of enrollment (see Table 4.1). Therefore, no significant differences exists

between the cocaine and non-cocaine exposed children in terms of their

sociodemographics at birth. At age three, still there are no statistically significant

differences between the two groups' race (z=-0.292; p=0.770) and household incomes

(z=-0.695; p=0.487).

Hypothesis 5: Using the presumed significance of race and class in the
social ecological model, I hypothesize that the families with a non-Black
child and higher household incomes will have children who perform better
on the behavioral and developmental scales than Black or non-Black
children living with impoverished families. In the multidimensional
model, I predict that the NonBlackAbove variable will be a stronger
predictor of behavior and development than the other RaceIncome dummy
variables given that BlackBelow is the reference group.

Summary

The basic hypothesis presented in this chapter is that, regardless of children's

cocaine exposure status, the postnatal environment is a significant predictor of behavioral

and developmental outcomes for pre-school aged children. The multidimensional model

summarizes the relationship between the prenatal, neonatal, and postnatal environment. I

predict that factors in the neonatal and the postnatal environment will compensate for the

risk associated with the prenatal environment, thus reducing the effects of prenatal

cocaine exposure on early childhood outcomes. I expect that in the bivariate analysis

presented in the following chapter, that the effects of prenatal cocaine exposure will be









minimal due to the strength of the other risks associated with this group of children. The

multivariate analysis is expected to indicate which biomedical and social ecological

factors measured in this study will explain differences in behavioral and developmental

outcomes.

In the next chapter, I test each of the above hypotheses within the context of the

multidimensional model. Table 3.1 lists the measures that are used to operationalize

each dimension in the model. The table outlines the constructs representing each

dimension and the standardized instruments used to measure each construct.















Table 4. 6: Description of Variables in Multidimensional Model
VARIABLES I DESCRIPTION I MEASURES
Dependent Variables (Scale)
Adaptive Behavior A composite score obtained from Vineland Adaptive Behavior Scales
(0-200) a behavior rating scale. (Sparrow, Balla, & Cicchetti 1984)
Development A composite score obtained from a Bayley Scales of Infant Development
10-2001 deelopmental rang scale I Ba\Ile\ lc)6)b
Prenatal Dimension
Cocaine Status Cocaine or non-cocaine exposed child Prenatally Exposed
No=0; Yes=l1 targeted for the research study. I
Neona"Ulmiennnsion
Infant Risk Factors Measure a newborns risk of poor health Hobel Neonatal (Hobel, Hyvarinen,
<10=risk outcomes. Okada, & Oh 1973; Hobel, Youkeles, &
Forsylhe 1979)
Postnatal Dimension
(Social Ecological Factors).. ... .
Microsystem
Psychological Measure of the primary Center for Epidemiologic Studies
Characteristics caregivers' risk for depression. Depression Scale (CES-D) (Radloff,
<16=risk 1977)
Mesosystem______________
Parenting Environment What are the nature, quality, and HOME Subscales II, IV, V, VI, VIII
(0-6) quantity of the interaction (Caldwell & Bradley 1984)
between the primary caregiver
~~~____~_______and the child in the home'?
Quality of Home The nature of the family's HOME subscales I, III, VII (Caldwell &
Environment physical residence, including Bradley 1984)
(0-6) assessments of safety, nurturing,
and developmental stimulation.
Exosystem
Family Support System The quality of support provided to Family Social Support Scale (Dunst,
(0-76) primary caregiver from others. Jenkins, & Tivette 1984) (Dunst, Trivette,
& Deal 1988)
Macrosystem
Sociodemographic Social and demographic Child's Race
Characteristics description of the primary Household Income
NonBlack=0; Black=l; caregiver and child.
NotPoor=0; Poor= I













CHAPTER 4
DATA ANALYSIS AND RESULTS




Analytical Procedures

In this chapter, I present the findings from the bivariate and multivariate analyses

of the multidimensional model. Each measure included in the model is presented in

terms of the differences between the prenatally cocaine exposed (PCE) and non-PCE

groups using the Mann-Whitney test. The Mann-Whitney test is a nonparametric test

equivalent to the t-test, but it does not require that the sample meet the strict assumptions

of a t-test. Nonparametric procedures are useful for studies like this one because of its

small non-normally distributed sample. The Mann-Whitney test assumes identical shapes

when comparing the population distributions, but it does not require the populations to be

"normal" or to be centered around the mean. The Mann-Whitney test is more powerful

than a median test since it uses the ranks of the cases. It requires an ordinal level of

measurement. The resulting mean rank sum is called U. The U represents the number of

times a value in the first group is smaller than a value in the second group, when values

are sorted in ascending order (Agresti and Finlay 1986; Hildebrand 1986; SPSS 2001).

In addition to analyzing the differences between the prenatally cocaine and non-

cocaine exposed infants on the measures of their neonatal and postnatal social

environments, the relationship between each environmental factor is analyzed using the

Spearman's correlation coefficient. Spearman is a nonparametric version of the Pearson








correlation coefficient. It is based on the ranks of the data rather than the actual values.

Like the Pearson's correlation coefficient, the values of the coefficient range from -1 to

+1 where the sign of the coefficient indicates the direction of the relationship. The

absolute value indicates the strength of the relationship with larger absolute values

indicating stronger relationships between variables (SPSS 2001).

The multivariate analysis includes the measures of the prenatal, neonatal and

postnatal environment as biomedical, behavioral and social predictors of behavior and

development. Using generalized linear regression which allows the response variable to

have a non-normal distribution, I have created three models of behavior and three models

of development. The standardized regression coefficients are presented in order to

demonstrate the relative strength of each predictor in each model. The higher the

standardized regression coefficient and the smaller the p-value, the stronger the predictor.

The first model includes the measure of prenatal risk or prenatal cocaine exposure as a

predictor of either behavior or development. The second model adds the neonatal factor

or the measure of health risks and the third model includes the prenatal, neonatal and the

postnatal factors as predictors of behavior and of development. These models reflect the

proposed additive or compensatory influence each factor has on pre-school aged

children's behavior and development.

The analysis begins with the questions posed in the introductory chapter. First, is

there a difference in the behavior or the development of PCE children and non-PCE

children? Second, to what extent do neonatal and postnatal factors hinder or enhance

behavioral and developmental outcomes among this group of "at-risk" children?








Differences between PCE children and non-PCE children

Birth Characteristics

In chapter 4, I described the characteristics of the longitudinal sample at birth.

The number of subjects used in this study has been reduced from 308 to 234, in order to

have complete data on each of the measures included in the multidimensional model.

The following table describes the characteristics of the reduced sample at birth.




Table 5. 1: Birth Outcomes by Cocaine Exposure Status, Means+/-Standard Deviations
(Mean Rank)___________________
Non-Cocaine Cocaine Total Sample Mann-
Exposed Exposed Whitney
_____n=119 n=115 N=234 p-values
Female 0.43+/-0.05 0.53+/-0.05 0.47+/-0.50 0.197
Infants (112.64) (122.53)_______________
Premature 0.08+/-0.28 0.13+/-0.34 0.11+/-0.31 0.252
Infant = Yes (114.83) (120.26)

Infant's Birth 3179.9+/-618.7 2985.5+/-614.2 3130.5+/-620.2 0.034
Weight (126.7) (108.0)

Hobel 44.62 +/-19.36 54.61 +/-17.47 49.53+/-19.08 0.000
Prenatal (124.47) (176.04)

Hobel Labor & 19.75 +/17.44 18.96 +/14.44 19.36 +/-16.01 0.789
Delivery (148.63) (151.40)

Hobel 15.43 +/-29.93 15.64 +/-21.46 15.54 +/-26.06 0.107
Neonatal (142.64) (157.51)



When compared to the longitudinal sample described in Table 4.2, the mean

percentage of females in the sample has not changed (47%), even though the significance

level has changed due to the smaller sample size. The difference in the number of

premature births for the PCE and non-PCE groups has changed. There are fewer








premature children in the reduced sample and the difference between the non-PCE and

non-PCE groups is no longer statistically significant (p=0.252). There is a statistically

significant difference between the two groups in terms of mean birth weight for the

children, however this sample indicates that the average birth weight of the children

included in the three year follow up is higher than the initial longitudinal sample.

The differences between the initial longitudinal sample and the children included

in the three year follow up are subtle, but they do indicate that the children with higher

birth weights and fewer perinatal complications were more likely to have completed all

assessments at age three. Unfortunately, this is a consequence of doing longitudinal

research with human subjects that is very difficult if not impossible to control. The

families that are most accessible are typically the ones who are engaged in research and

from these numbers those families tend to be the ones with stronger or healthier children,

a fact that is true for most research, but is often negated. The researchers made a special

effort to control for the social demographics and health related risks that may confound

any observed differences between the PCE and non-PCE children, but it is difficult to

control for attrition. Therefore, it is important to keep in mind that the conclusions

drawn in the following analyses reflect the differences between the two groups of the

families who were engaged in this research project, but they may or may not reflect the

general population.

Neonatal Environment

Maternal drug use during pregnancy is one of the items used to construct the

Hobel Perinatal Complications Scale. It is expected that children who were prenatally

exposed to any type of illicit drug or alcohol will have more health problems than their

non-exposed counterparts. The neonatal assessment represents the risk of health








complications for the newborn infant, while the prenatal score is an assessment of

prenatal risk to the fetus and the labor and delivery score is used to determine the risk of

complications that the mother and child may encounter during delivery. Only the

neonatal assessment score is used in the multivariate analysis, because it encompasses the

prenatal and the labor and delivery risk scores. In this sample of "at-risk" children, there

is no statistically significant difference between PCE children and the non-PCE

children's mean score on the neonatal complications scale (z=-1.611; p=0.107).

The neonatal score of the PCE children (mean=15.6; standard deviation=21.5) is

slightly higher than the non-PCE children (mean= 15.4; standard deviation=29.9), but this

difference is only approaching statistical significance, whereas the difference between the

PCE (mean=54.6; standard deviation=17.5) and non-PCE children (mean=44.6; standard

deviation=19.4) on the prenatal risk assessment is significantly different (p=0.000).

Hence, as time passes, among the fetuses that survive the pregnancy and the infants who

survive the birthing process, the health risk associated with prenatal cocaine exposure

dissipate. But, it is important to note that both groups have neonatal risk scores above 10,

which indicates considerable risk for health complications during infancy and early

childhood.


















i- l i


50-


40- Hobel Prenatal


30,
Hobel Labor &
Delivery

20,
Mean *Hobel Neonatal
10i
Non-PCE PCE
n=119 n=115
N=234


Figure 5. 1: Hobel Complications Scale by Prenatal Cocaine Exposure Status



Postnatal Environment

Within the postnatal environment, prenatal cocaine exposure has even less

association with the nature of the environment. As shown in Tables 5.2 and 5.3, there are

minimal differences between the PCE and non-PCE children in terms of their social

environment as measured using the social ecological factors.








Microsystem. According to the results on the Center for Epidemiologic Studies-

Depression scale (CES-D), there is a statistically significant difference between the

depression status of the primary caregivers with PCE children and those with non-PCE

children. The primary caregivers with non-PCE children have slightly higher scores on

the CES-D (mean = 22.88; standard deviation=9.17) than the primary caregivers with

PCE children (mean = 19.97; standard deviation = 10.77). Both groups received scores

above 16, which typically indicates a risk for depression. On average, both groups of

caregivers reflect a moderate risk for depression, a trend that is also related to whether or

not the primary caregiver is the biological mother or not. Non-cocaine exposed children

are more likely to be living with their biological mother than cocaine exposed children

(chi-square = 75.97, p=0.000). This relationship between prenatal cocaine exposure and

parental status is explored further in the section of this chapter that describes the

relationships between the environmental factors and outcomes.

Mesosystem. The nature of the home environment was measured using the

HOME scale. If one were to assume that women who used drugs during their pregnancy

are less likely to live in and/or maintain stable and nurturing households due to their drug

behavior, then it would be safe to postulate that the children who were prenatally exposed

to cocaine would live in poorer home environments. However, among the children in this

sample, very few of the PCE children still live with their biological mother at age three.

So, this hypothesis can not be fully evaluated.








Table 5. 2: Postnatal Differences between PCE and non-PCE children at the
Microsystem and Mesosystem Level, Mean +/- Standard Deviation (Mean Rank)
Non-Cocaine Cocaine Total Mann
Exposed Exposed Whitney
N=119 N=115 N=234 p-value
Microsystem Psychosocial Dimension
CES-Depression Scale 22.88+/-9.17 19.97+/-10.77 21.45+/-10.07 0.018
1 (127.78) (106.87)
Mesosystem Home Environmental Dimension______
Physical Environment 5.63+/-1.57 5.71+/-1.65 5.67+/-1.60 0.856
(116.71) (118.32)______
Home I: 5.61+/-2.63 6.11+/-2.75 5.86+/-2.70 0.222
Learning Stimulation (112.22) (122.97)
(0-11)______
Home III: 5.46+/-1.81 5.25+/-1.87 5.36+/-1.84 0.310
Physical Environment (121.76) (113.10)
(0-7)_____________________
Home VII: 5.81+/-1.68 5.76+/-1.62 5.78+/-1.65 0.755
Variety in Experience (118.83) (116.12)
(0-9)__________________________
Parenting Environment 4.59+/-0.96 4.70+/-0.10, 4.26+/-0.89 0.604
(115.79) (119.27) _____....,:;_,
Home II: 6.26+/-1.01 6.30+/-1.04 6.28+/-1.02 0.579
Language Stimulation (115.34) (119.73)
(0-7)____________________________
Home IV: 5.17+/-1.55 5.49+/-1.65 5.32+/-1.60 0.041
Warmth and Acceptance (108.82) (126.48)
(0-7)____________________________
Home V: 3.73+/-1.26 4.06+/-1.08 3.89+/-1.18 0.058
Academic Stimulation (109.64) (125.63)
(0-5)__________________________
Home VI: 3.22+/-1.66 2.95+/-1.31 3.09+/-1.24 0.117
Modeling (124.13) (110.64)
(0-5)__________________________
Home VIII: 2.82+/-1.20 .65+/-1.12 2.74+/-1.16 0.131
Acceptance (123.81) (110.97)
(0-4) ______________________________ ______


There are no significant differences between the PCE and non-PCE children's

parenting environments or their physical home environments, with one exception. The

parenting environment of the PCE children is slightly higher on the measure of maternal

warmth and acceptance (p = 0.041) and academic stimulation (p = 0.058). These








differences indicate that the prenatally cocaine exposed children are living with primary

caregivers who show more warmth and acceptance and provide more academic

stimulation than the primary caregivers of the non-PCE children. The fact that most of

the primary caregivers for the PCE children are not the biological mothers makes this

finding quite interesting. Yet, further analysis did not show a relationship between the

different subdomains on the HOME and the caregiver status of the biological mother.

Across the different domains of the HOME, the physical home environment and the

parenting environment of the PCE and non-PCE groups is not significantly different, nor

is it different when controlling for the caregiver status of the biological mother.

Exosystem. The exosystem is measured using the Family Social Support Scale.

This scale assesses the quality of the social support available to the primary caregiver.

Within this study sample, the primary caregivers with a non-cocaine exposed child had a

mean level of support of 39.31 (standard deviation = 10.93) and the primary caregivers

with a prenatally cocaine exposed child had a slightly lower level of support at 37.52

(standard deviation = 12.92). These differences were not statistically significant

(p=0.112), but they do represent a relatively low level of social support. On a scale of 0

to 76, the mean level of support for both groups was only 38.43 with a standard deviation

of 11.96.

Due to how the social support scale is defined in different studies, it is difficult to

ascertain what a "normal" level of support is for "at-risk" families. The range of support

varies from one population to another depending on how many supports are available to

the family, but a mean that falls at or below the midpoint of 38 indicates that there is








either a low level of support available to these families and/or the available supports are

not necessarily helpful (see Table 5.3).


Table 5. 3: Postnatal Differences between PCE and non-PCE children at the Exosystem
and Macrosystem Level, Mean +/-Standard Deviation (Mean Rank)
Non-Cocaine Cocaine Total Mann-
Exposed Exposed Whitney

n=119 n=115 n=234 p-value
Exosystem Family Social Support Dimension
Family Support Scale 39.31+/-10.93 37.52+/-12.92 38.43+/-11.96 0.112
(12.00 76.60) 1 (124.40) (110.36)
Macrosystem Social Structural Dimension
Child's Race=Black 0.83+/-0.38 0.82+/-0.39 0.82+/-.38 0.770
(118.31) (116.63)________________
Families Living Below 0.71+/-0.45 0.65+/-0.48 0.68+/-.47 0.308
Poverty (121.07) (113.80)________________
Black Below Poverty 0.68+/-0.47 0.60+/-0.49 0.64+/-0.48 0.199
~~____~______(122.14) (112.70)__________________
Black Above Poverty 0.15+/-0.36 0.22+/-0.41 0.18+/-0.39 0.193
~~____~______(113.70) (121.43)________________
Non-Black Below 0.03+/-0.18 0.05+/-0.22 0.04+/-0.20 0.484
Poverty (116.43) (118.60)_________________
Non-Black Above 0.13+/-0.34 0.13+/-0.34 0.13+/-0.34 0.928
Poverty (117.73) (117.26)________________


Macrosystem. According to the tenets of urban sociology, poverty causes social

disadvantages that can have a profound effect on the victim's health, behavior, and life

course trajectory (Furstenberg, Cook, Eccles, Elder, and Sameroff 1999; Wilson 1987).

By matching the two groups, the researchers were better able to control for the

differences in birth outcomes caused by the cumulative effects of social disadvantage.

There are no significant differences between the PCE children and the non-PCE children

in terms of their race or household income by design. The majority (82%) of the children








in the sample are classified as Black and 68% of the families live below the poverty line

with an income of less than $16,600 for a family of four.

Sixty-four percent of the children are Black and living below poverty. Eighteen

percent of the sample includes black children living above poverty. Non-Black children

living below poverty make up 4% of the sample, while non-black children living above

comprise 13% of the sample. Respectively, the majority of the Black children live below

poverty and the majority of non-Black children live above poverty.

Thus far, there have been very few differences between the PCE and the non-PCE

children in this sample. The data indicates that there are slight differences in the

characteristics of the home and the primary caregivers' psychosocial status; whereas the

cocaine exposed children are residing with caregivers who are less likely to be depressed

and who express more warmth and acceptance towards the child. Given the fact that the

majority of the PCE children are not living with their biological mother at age three, these

findings are interesting, but not antithetical. The better home environment among the

PCE children suggests that the PCE children have not only been removed from their

biological mothers, but they have been removed from less supportive environments as

well. The findings presented thus far also indicate that the group of children included in

this sample have considerably high neonatal risk scores, a greater likelihood of living

with a depressed primary caregiver and in homes that can be considered less than perfect.

Behavior and Development

The next question asks whether there is a relationship between the child's

environment and the child's behavioral and developmental outcomes. Despite the

contradictory and inconclusive evidence in the literature on the effects of prenatal cocaine

exposure, most people assume that children who have been prenatally exposed to cocaine








are developmentally delayed resulting in severe behavior problems. The research

findings presented here suggest otherwise.

Among this sample of 234 children, there are no significant differences between

the PCE children and the non-PCE children on any of the behavior or developmental

domains measured by the Vineland Adaptive Behavior Scales (Vineland) or the BiY.le\

Scales of Infaint Development (Bayley). Table 5.4 shows the results for both scales and

each of the subdomains. In fact, the PCE children actually have higher mean scores than

the non-PCE children on the Vineland's communication domain (p = 0.151),

socialization domain (p = 0.091), and motor skills domain (p = 0.169). Even though

these results do not represent a statistically significant difference, they are indicative of

the trends mentioned earlier, whereas the PCE children are living with less depressed

caregivers, in more nurturing home environments, with slightly higher household

incomes than non-PCE children.

However, as indicated in Table 5.4, the scores for both groups on the Vineland

(mean = 98.5; standard deviation =14.7) and the Bayley (mean = 90.4; standard deviation

= 15.4) are teetering between moderately low and average given the high standard

deviation. Fifty-five percent of the children scored below 100, the national average, on

the Vineland, while 72% received scores below the national average of 100 on the

Bayley. The low performance on these nationally standardized tests indicate that this

group of children, regardless of their cocaine exposure status is at risk of poor behavioral

and developmental outcomes. This finding is inconsistent with the popular belief about

the effects of cocaine exposure on behavior and development, but it is consistent with the

belief that this group of children is at risk of developmental problems which could lead to








behavioral difficulties, for there is a significant correlation between the behavior and

development (rho=0.473; p=0.000).




Table 5. 4: Adaptive Behavior and Development by Cocaine Exposure Status, Means +l-
Standard Deviations (Mean Rank)________
Non-Cocaine Cocaine Total Mann-
Exposed Exposed Whitney

n=119 n=115 N=234 p-values
Adaptive Behavior 97.27 +/- 14.19 99.81 +/- 15.16 98.51+/- 14.70 0.314
(Vineland) (113.12) (122.03)
(Range: 20 -160)____________________
Communication 98.22+/-14.32 101.20+/-13.99 99.69+/- 14.20 0.151
110.74 123.43_______________
Daily Living Skills 103.23+/-14.05 101.63+/-14.77 101.52+/- 0.515
120.33 114.57 13.79 ______
Socialization 97.68+/-12.62 99.92+/-13.09 98.63+/- 12.06 0.091
110.14 125.11_____________
Motor Skills 95.32+/-16.13 96.50+/-15.39 96.05+/- 15.62 0.169
111.52 123.69_____________
Development 90.59+/- 15.97 90.21+/- 14.88 90.40 +/-15.41 0.624
(Bayley) 119.63 115.30
(Range: 50 150)______________________________
Motor Skills 95.95+/-18.15 94.07+/-18.06 95.02 +/-18.09 0.354
121.53 113.33_______________
Mental Skills 85.23+/-16.79 86.35+/-15.62 85.78 +/-16.20 0.762
116.18 118.86______ _________
Interpretation of Scales: Low (> 69), moderately low (70-84), adequate (85-115),
moderately high (116-130), high (131 or above)


So, if prenatal cocaine exposure does not explain the differences between children

in terms of their behavior and development, what does? The multidimensional model

suggests that the neonatal and postnatal environments compensate for the risk caused by

the prenatal environment. In this next section, I explore the relationship between each

environmental dimension, behavior and development, and the early childhood outcomes.








Environmental Factors and Early Childhood Outcomes

Prenatal Risk and Outcomes

In the bivariate analysis, PCE was not significantly correlated with any of the

social ecological factors from the postnatal environment with the exception of the

primary caregiver's depression score (rho= -0.155; p=0.018). Apparently, the primary

caregivers who were caring for a non-PCE child had higher scores on the depression

inventory than the primary caregivers with a PCE child. This enigmatic finding may be

better understood within the context of the relationship between the primary caregiver

and the child.

Approximately 50% of the PCE children did not live with their biological mother

at age three, whereas only 4.2% of the non-PCE did not live with their biological mother

at this same point in time. Hence, the non-cocaine using mothers are more likely to

maintain custody of their children (rho=-0.514; p=0.000), and they are more vulnerable to

depressive symptomology (rho=0.371; p=0.000) which is significantly correlated with the

nature of the parenting environment (rho = -0.179; p=0.006), the physical environment

(rho=-0.157; 0.017), family social support (rho=-0.185; p=0.005); and the family's

poverty status (rho=-0.143; p=0.029) as shown in Table 5.5. Therefore, regardless of

cocaine exposure status, the children in this study are "at-risk" of living in poor home

environments with low quality social support, and limited economic resources when

compared to the norms used in the standardization tables for each measure.

Neonatal Risk and Outcomes

Using the multidimensional model, I predict that there will be a relationship

between the neonatal environment and behavior and development. The model infers that

positive health outcomes for newborns results in better behavioral and developmental








outcomes for toddlers. While prenatal cocaine exposure poses a risk to behavioral and

developmental outcomes, the neonatal and postnatal environment serve to compensate for

prenatal complications. "At-risk" children who are born in a hospital, typically receive

extra care and attention neonatally; and in the case of the cocaine exposed children, the

extra care extends beyond their stay in the hospital. This is exemplified by the

termination of parental rights for drug addicted mothers and/or the child's enrollment in

social service programs designed to monitor the child's progress (i.e. Department of

Children and Families). Hence, according to the compensatory model of resilience used

to organize the multidimensional model, low neonatal health risk and a positive postnatal

environment should compensate for the risk associated with a negative prenatal

environment.

In this sample of "at-risk" children, the neonatal risk score is significantly

correlated with adaptive behavior (rho=-0.143; p=0.028) and it is marginally correlated

with development (rho=-0.110; p=0.094) as shown in Table 5.5. As the number of health

complications experienced by newborns decreases, the scores on the Vineland and the

Bayley increase. In other words, children with fewer neonatal health complications tend

to have better adaptive behavior at age three; whereas this is true for development but the

relationship is not as strong.

In terms of the relationship between the neonatal and postnatal environments,

there is no significant relationship between the two. In the bivariate analysis, each of the

three dimensions is independent of the other with one exception. A child's score on the

neonatal health complications assessment is significantly correlated with the families'

poverty status (rho=-0.132; p=0.044). The children with the greater number of neonatal








complications at birth are more likely to be living with a family that's average income is

above the poverty line. Again, we encounter the relationship between risk and improved

surroundings. The children with the greatest risk at birth, whether it was from prenatal

cocaine exposure or some other contributing prenatal factor, appear to be living in better,

less impoverished environments by the time they reach age three.

Postnatal Factors and Outcomes

In this section, I discuss the hypotheses pertaining to the social ecological model

that represents the postnatal environment described in Chapter 4. It is presumed that the

non-Black children living above poverty with psychologically stable caregivers who are

more nurturing and who have better quality home environments and social support

networks will exhibit better behavioral and developmental outcomes. This hypothesis is

true to some extent. With the exception of the primary caregiver's psychosocial status,

each dimension in the postnatal environment is significantly correlated with both

behavior and development. The primary caregiver's level of depression is significantly

related to development, but not to behavior. The children with lower developmental

scores, tend to live with primary caregivers who report a higher level of depressive

symptoms (rho=-0.129; p=0.049). The lack of a relationship between depression and

behavior and the existing relationship between depression and development is quite

interesting, but difficult to interpret due to the bi-directional nature of parent-child

interaction. Do "problem" children cause caregivers to experience the symptoms of

depression or does depression impair the caregivers' parenting capabilities? Further

consideration is given to this dilemma in the subsequent chapter.








Table 5. 5: Correlation between Adaptive Behavior and Development, Prenatal Cocaine
Exposure, and Environmental Factors, Spearman Correlation Coefficient, n=234
Prenatal Behavior Development Neonatal
Cocaine (Vineland) (Bayley) Health (Hobel)
Exposure______________________
Prenatal 1.000
Cocaine
Exposure
Behavior .066 1.000 --
(Vineland) ____________ ______
Development -.032 .473*** 1.000
(Bayley)_______________
Neonatal Risk .106 -.143* -.110 1.000
Assessment_________________
PCG Depression -.155* -.071 -.129* -.085
Scale_________________ ____________________
Physical .012 .324*** .237*** .058
Environment________________
Parenting .026 .387*** .276*** -.062
environment________________
Family Social -.104 .160** .174** .018
Support______________
Child's Race -.019 -.199** -.256*** .059
Family below -.067 .000 -.1090 -.132*
poverty_____________________
Black Below -.084 -.022 -.134* -.085
Poverty_______________
Black Above .085 -.167** -.085 .163*
poverty_______________
Non-Black .046 .054 .066 -.102
Below Poverty__________ __________ ____________________
Non-Black -.006 .191** .247*** -.005
Above Poverty__________ __________ ____________________
_p< .10; *p< .05; **p<.l01; ***<001 (2-tailed test)


The mesosystem is characterized by the nature of the physical home environment

and the parenting environment. According to the correlations shown in Table 5.5, there

is a statistically significant relationship between the parenting and physical environments

and behavior and development. Children with better behavioral and developmental

outcomes tend to live in homes with more conducive parenting and physical








environments. The correlations for these measures, though statistically significant differ

on the behavior and the development scales. The behavioral score has a higher

con-rrelation with the parenting (rho=0.387; p=0.000) and the physical (rho=0.324;

p=0.000) environment than the development scale (rho=0.276; p=0.000; rho=0.237;

p=0,000) respectively. It is possible that these small differences are due to the nature of

the assessments, in that the behavioral assessment is conducted with the primary

caregiver and an autonomous clinician conducts the developmental assessment.

One might argue that the primary caregiver's assessment of the child and of the

nature of their home environment is biased resulting in socially desirable scores and

stronger correlations between the measures that are dependent on the caregiver's

assessment. Yet, this argument does not hold true when looking at the relationship

between the depression score and outcomes and the family social support score and

outcomes. In both cases the correlations with the Bayley, which is determined by the

clinician, are higher than the correlations with the Vineland, the interviewer and

caregivers joint assessment of the child's behavior. Therefore at this point it is unclear

why the correlations differ.

The behavioral and the developmental assessments of children living in homes

with a better quality of support are better than the assessments of children living in homes

with less social support. Here, the developmental score correlates with the family social

support scale at rho = 0.174 (p = 0.008) while the behavioral score is still significantly

correlated but has a smaller coefficient (rho = 0.160) and higher p-value (p = 0.014).

At the macrosystem level, the socioeconomic status of the family is related to the

family's level of social support and to the nature of the home environment (see Table








5.5). Families living below the poverty level received lower scores on the physical

measure of the home environment (rho=-0.396; p=0.000), on the parenting dimension

(rho=-0.209; p=0.001) and on the family social support scale (-0.135; p=0.039). Of the

dummy variables created to classify the families' social and economic background, the

non-Black families living above poverty had the strongest correlation with the outcome

measures. Non-Black children living above poverty are more likely than the other groups

of children to receive higher scores on the Vineland behavioral assessment (rho=0.191;

p=0.003) and on the Bayley developmental scale (rho=0.247; p=0.000). In contrast,

Black children have poorer developmental outcomes when they also live in poverty

stricken households. Yet, Black children who live above the poverty line tend to score

lower on the behavioral scale. The race-class effect observed here shows the complexity

of the relationship between socioeconomic status and outcomes. This relationship is

explored further in the subsequent chapter.











Table 5. 6: Complete Correlation MatrixSpearman Correlation Coefficients, n=234
= =, ,- -- -n- --
(1) Cocaine 1.000
Exposure
(2) Behavior .066 1.000
(Vineland)
(3) l)evp'mt -.032 .473* 1.000 .. .. ...
(Bayley) _______
(4) Neonatal .106 -143* -.110 1.000 ...
Risk (Hobel)
(5) PCG -.155* -.071 .129* -.085 1.0L00 .. .
Depression
(6) '' 012 324* 237* -.001 -.157* 1 000 .- .- .- --
Environment __________ _____
(7) Parenting .026 .387* .276* -.062 -.179* .568* 1.000 .- --.. ..
Environment
(8) Family Social -.104 .160* 174* .018 -.185* .204* .081 1.000 .. ..
Support ________ _______
(9) Child's Race -.019 -.199* -.256* .059 -.007 -.437* -.206* .065 1.000 .

(10) Family -.067 .000 -.109 -.132* .143* -396* -.209* -. 135* .436* 1.000 -....
Below Poverty
(11) Black -.084 -.022 -.134* -.085 .118 -421* -.208* -.094 .616* 909* 1.000
Below Poverty
(12) Black .085 .167* -.085 .163* .154* .093 .055 .180* .219* -.698* -.634* 1.000
Above Poverty _____
(13) Non-Black .046 .054 .066 -.102 047 .090 .012 -.089 -.458* 144* -.282* -.100 1.000
Below Poverty_____ __________ _____ _____ _______ _____ _______ _____ __
(14) Non-Black -.006 .191* .247* -.005 .020 .436* .223* -.020 -.848* -.575* .522* -.185* -.083 1.000
Above Poverty I

*p<0.05 (2-tailed test)









Multidimensional Model of Behavior and Development

The primary hypothesis presented in this study is that, regardless of children's

cocaine exposure status, the postnatal environment is a significant predictor of behavioral

and developmental outcomes for pre-school aged children. This hypothesis is based on a

compensatory theory of resilience and Bronfenbrenner's ecological theory of human

development (see Chapter 3 for more detail). Based on the results from the bivariate

analysis, it is clear that the neonatal and postnatal environment are significantly related to

children's outcomes, whether they have been prenatally exposed to cocaine or not. In the

following section, I use multivariate analysis to examine the relative contributions of

biomedical, behavioral, and social factors on early childhood outcomes.

For the sake of simplicity, behavior and development have been examined

comparatively throughout the bivariate analysis. However, behavior, as a construct is

different from development. According to Ribes (1996), behavior has been used to

explain the process of development, while development has been defined as the

qualitative changes in behavior. The concept of behavioral development was introduced

in the 1950s under the rubric of social learning theory. Since that time, theories about the

behavioral and developmental processes have been developed and tested by sociologists,

behaviorists, psychologists, biologists, ecologists, and so on (see Bijou and Ribes 1996

and Lerner 1983 for historical perspective). Roberts, Maddux, and Wright (1984:57)

define the process of development as a process of systematic and measurable changes in

behavior that occur "across the life span- from the prenatal period through life to death."

In this vein, this research explores how prenatal, neonatal, and postnatal changes impact

behavior and development, two processes that are significantly correlated. In this








research, I do not contend that development causes behavior, but the relationship between

behavior and development is definitely worth exploring in future studies of "at-risk"

children.

Behavior

In this study behavior refers to the intentional or unintentional actions exhibited

by children as they learn how to adapt to their surroundings. The Vineland Adaptive

Behavior Scale specifically measures the children's ability to demonstrate skills and to

perform tasks that are necessary components of social interaction and individual survival.

The behavior of children who have been prenatally exposed to cocaine has been of

particular interest to those who care for these children at home and in public institutions

because of the presumed behavioral problems associated with prenatal cocaine exposure.

The early childhood behavior of this group of children is assumed to be a proxy for later

life behaviors that may be problematic.

Among the children included in this study, prenatal cocaine exposure does not

appear to be related to behavior. There are no significant differences between the PCE

children and non-PCE children. Indeed the bivariate regression model is not significant.

In model 2, adding a second biomedical component (i.e. neonatal risk) makes the model

significant. Higher neonatal risk of physiological complications is associated with lower

behavioral scores (see Table 5.7). Prenatal cocaine exposure is not correlated with

adaptive behavior and it is not predictive of behavior in the final multivariate model. In

Table 5.7, the multidimensional model (Model 3) is compared to the biomedical model

(Model 2). Adding social contextual factors significantly improves the models fit and the

explained variance.








The strongest predictor of adaptive behavior is the mesosystem, specifically, the

nature of the parenting environment (Beta =0.278; p= 0.000). Adaptive behavior scores

are highest among children who have a positive parenting environment. The neonatal

environment is the second strongest predictor in the multidimensional model of behavior.

Children who had greater neonatal risk tend to have lower behavioral scores three years

later (Beta =-0.172; p=0.004). Last, Model 3 indicates that Black children living above

the poverty line have poorer adaptive behavior than the below-poverty Blacks (Beta=-

0.165; p=0.009). It is interesting to note that non-Blacks living above and below poverty

have more similar behavior scores than poor Black children.

In the bivariate analysis, the families' level of social support and the quality of the

physical home environment were significantly correlated with behavior, but they did not

maintain their strength in the multivariate analysis. In general, the neonatal environment

and aspects of the postnatal environment are significant predictors of behavior regardless

of cocaine exposure status (Incremental F=7.292; p<0.001).








Table 5. 7: Multidimensional Models of Adaptive Behavior and Development,
Standardized Regression Coefficients (p-values), n=234
Vineland Adaptive Behavior Bayley Scales of Infant
Scales T 3Development
Variables 1 2 3 4 5 6
PRENATAL RISK FACTOR
Cocaine 0.087 .087 0.093 -0.012 -0.011 -0.019
Exposed (0.187) (0.173) (0.117) (0.851) (0.860) (0.755)
NEONATAL HEALTH FACTOR
Hobel -0.218 -0.172 -0.268 -0.254
Neonatal (0.001) (0.004) (0.000) (0.000)
POSTNATAL SOCIAL FACTORS_____________________
Depression -0.022 --0.108
Scale ____________(0.715) ____________(0.089)
Parenting 0.278 -0.175
Environment (0.000) ____________ (0.022)
Physical 0.129 --0.013
Home (0.120) (0.877)
Environment_____________________________
Family 0.093 -0.140
Social (0.133) (0.030)
Support_______________________________
Black Above .- -0.165 --0.044
Poverty (0.009) ____________ (0.499)
Non-Black -- 0.037 -0.065
Below (0.534) (0.299)
Poverty
Non-Black 0.071 -- -0.181
Above (0.292) (0.010)
Poverty__________________________________
R2 0.007 0.055 0.254 0.000 0.072 0.194

F 1.751 6.723 8.495 0.035 8.974 5.993
(0.187) (0.001) (0.000) (0.851) (0.000) (0.000)
Incremental Between Between
F Model 7.292 Model 4.8436
_____3&2 ______6&5_______


Development

Like behavior, there are numerous ways of measuring development.

Development can be defined in broad or narrow terms. In this study, the Bayley Scales of

Infant Development are used to measure the development of each child's mental and








motor skills. Development refers to children's achievement of particular growth

parameters in relation to their chronological age. Development is the process of growing

and reaching physiological maturity. From a biomedical standpoint, development is a

physiological process, but when placed within the context of the social environment, the

physiological processes can be directly or indirectly influenced by external factors.

The Bayley measures the child's fine and gross motor skills, abstract reasoning,

memory, learning and problem solving abilities. These domains reflect

neurodevelopmental functioning, an area that can be compromised by the injection of

cocaine (see Lauder 1991; Spear, Kirstein, Frambes 1989; Dow-Edwards 1991).

According to Mayes and Bornstein (1995:253):

Cocaine influences brain development directly through effects on developing
neurotransmitter systems critical to neuronal differentiation and brain structure
formation and indirectly through effects on blood flow to the developing fetal
brain... Cocaine blocks the reuptake of dopamine, norepinephrine, and 5-HT by
the presynaptic neuron [Swann, 1990], a process that is primarily responsible for
inactivation of neurotransmitters.

There are certain parts of the brain that are insensitive to the effects of cocaine,

but in fetal brain development, the effect cocaine has on the monoaminergic

neurotransmitters is critical to brain structure and neuronal formation. Therefore,

prenatal cocaine exposure places children at risk of poor neurodevelopmental

functioning. Poor neurodevelopmental functioning can affect children's ability to react to

stimuli, modulate arousal, and regulate their attention (Mayes and Bornstein 1995).

In the multivariate analysis shown in Table 5.7, prenatal cocaine exposure is not a

significant predictor of development in any of the models. Similar to the behavior

models, the first model (Model 4) is unremarkable. Model 5 shows the impact of adding

in an additional biomedical factor. The model itself is significant and it explains 7.2% of








the variance compared to Model 4 where prenatal cocaine exposure does not explain any

of the variability in development. Similar to the behavior model, children with higher

neonatal risk tend to have lower developmental outcome scores. The final model, in

which the social factors are added, is a significant improvement over the biomedical

model (Model 5). The strength of the neonatal risk score changes very little when the

postnatal factors are added. The addition of postnatal environmental factors enhances our

understanding of development by explaining nearly 20% of the variance in development,

which is nearly three times more than what the biomedical factors explained alone.

Of the postnatal factors included in the multidimensional model, there are

significant effects from the micro-, meso-, exo-, and macro dimensions of the social field.

At the microdimension, children whose primary caregivers have higher levels of

depression are marginally more likely to have lower developmental scores (Beta = -

0.108; p = 0.089). At the mesodimension, again the parenting environment is associated

with developmental outcomes (Beta = 0.175; p = 0.022). However, the physical home

environment, which was correlated with development in the bivariate analysis, is no

longer significant in the multivariate analysis (Beta = -0.013; p = 0.877). In contrast to

the behavior model, family social support is significantly associated with development

(Beta = 0.140; p = 0.030). Specifically, children with higher development scores live

with caregivers who have a more helpful social support network. Non-Black children

living above poverty have significantly higher scores on the Bayley than the poor Blacks

(Beta = 0.181; p = 0.010). Both poor non-Blacks and non-poor Black children have

developmental scores similar to poor blacks. This means that social class shapes








developmental score differences among non-Blacks, but not among Blacks. The macro

dimension also works differently for behavioral and developmental outcomes.


Summary

In conclusion, each dimension in the multidimensional model explains some

portion of the variability in development and in behavior. The neonatal environment is

the strongest predictor of outcomes, while the postnatal environment adds to our

understanding of behavioral and developmental differences among children who are

considered to be "at-risk" of early childhood delays. In the multivariate analysis, poor

neonatal health and postnatal social and economic disadvantage place these children at

greater risk for poor behavioral and development outcomes, while fewer neonatal

complications and a positive home environment full of social support compensates for

those risks.

The fact that there are no statistically significant differences between the children

who were and who were not prenatally exposed to cocaine does not mean that children

who have been prenatally exposed to cocaine are not "at risk." As indicated by their

mean scores on the behavior and development scales, on average the PCE and non-PCE

children from this at-risk sample scored below normal (< 100) on both scales. The lack of

a significant difference between the two groups does indicate that factors other than

prenatal cocaine exposure either help or hinder the development of positive early

childhood outcomes. The sample of children in this study represent a group of children

who are at risk of poor outcomes due to socioeconomic disadvantage, which is related to

familial instability, poor parent-child interaction, maternal depression, and insufficient

family and social support. The problems associated with prenatal cocaine exposure,





87


such as lower birth weight, perinatal health complications, and prematurity were not

profound.

Again, the social and environmental factors considered in the multidimensional

model were stronger predictors of outcomes for children at age 3 than their risk of

prenatal cocaine exposure. In the concluding chapter, I expand upon these findings and I

discuss their social and political implications for prenatally cocaine exposed children,

their biological mothers, and the many people who care for them.














CHAPTER 5
RISK AND RESILIENCY IN EARLY CHILDHOOD


Early Childhood Risk

Prenatal Cocaine Exposure

The goal of this research has been to identify the relative influence of the prenatal,

neonatal, and postnatal environments on the behavior and development of pre-school

aged children. The children included in this study have been deemed "at-risk" because of

their prenatal cocaine exposure. Specifically, this research has been designed to

determine the effect that prenatal cocaine exposure, a prenatal risk factor, has on early

childhood outcomes. As a whole, literature on prenatally cocaine exposed children is

inconclusive due to contradictory results reported in the biomedical and behavioral

literature.

The biomedical literature has taken on a teratogenic approach that focuses on the

physiological effect cocaine has on the fetus and developing child. This problem-seeking

approach presupposes that cocaine exposure has a direct organic effect on prenatally

exposed children that will negatively influence later stage development. The biomedical

model is limited because it typically fails to consider the social and environmental factors

that can enhance or hinder behavior and development. The behavioral studies have tried

to compare the behavioral and developmental outcomes for PCE and non-PCE children,

which is difficult because of the methodological issues surrounding the study of prenatal