 

 

"ﬁr

v‘¢1\~\

.
.

 

 

THESHs

éL

Illi'lllli'lllllllllllllillllllllllfalllilll

3 1293 01579 9756

 

 

 

 

 

LIBRARY
Michigan State
University

 

 

 

This is to certify that the

thesis entitled

A Measure of Functional Impairment in Children
and Adolescents and its Reiation to
Symptomatology and Diagnosis

presented by

Jennifer S. Paul

has been accepted towards fulﬁllment
of the requirements for

M.A. degree in Psychology

 

 

AMA/VAL M
Majowfessor

Date April 25, 1996

 

0-7 639 MS U is an Affirmative Action/Equal Opportunity Institution

 

____ _. ___.__-———_-—_’—~——‘—-——————-—‘——‘-_-_

PLACE II RETURN BOX to remove thie checkout from your record.
TO AVOID FINES return on or More due due.

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MSU I. An Afﬂrrnetive WW Opportunity "Dilution
WM”

 

 

A MEASURE OF FUNCTIONAL IMPAIRMENT m CHILDREN AND
ADOLESCENTS AND ITS RELATION To SYMPTOMATOLOGY AND
DIAGNOSIS
By

Jennifer 8. Paul

A THESIS

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

MASTER OF ARTS

Department of Psychology

1996

 

ABSTRACT
A MEASURE OF FUNCTIONAL IMPAIRMENT IN CHILDREN AND
ADOLESCENTS AND ITS RELATION TO SYMPTOMATOLOGY AND
DIAGNOSIS
By

Jennifer S. Paul

The purpose of this study was to validate the Functional Impairment Scale for
Children and Adolescents (FISCA; Frank & Paul, 1995) by establishing its concurrent,
discriminant, and predictive validity. As such, the study also assessed the degree of
overlap between functional impairment, as measured by parent report on the FISCA,
and patient symptomatology and diagnoses. The total sample consisted of a) 400
children and adolescents inpatients (ages 6-17) and /or their parents in two Midwestern
psychiatric hospitals as well as b) 100 parents of patients from two outpatient clinics
associated with each of the hospitals. Correlations between the FISCA and other
measures of functional impairment and between the FISCA and parent and child reports
of children’s symptomatology supported the concurrent validity of the FISCA. FISCA
scores were able to discriminate between outpatients and inpatients; between patients
with no, low and high comorbidity; and in most cases, between patients with and
without particular diagnoses. The predictive validity of the FISCA was supported by
its ability to predict length of stay. Overall, the study showed that the FISCA is a
valid measure of functional impairment and can be a useful addition to traditional
assessment protocols that focus only on symptoms and diagnoses in assessing

psychological difﬁculties in children and adolescents.

 

ACKNOWLEDGMENTS

I would like to thank my thesis chair, Susan Frank, for her valuable guidance
and support at every phase of this project. I would also like to thank the other
members of my committee, Alytia Levendosky, Michael Lambert, and Gary Stollak for
their support and constructive feedback. In addition, I would like to thank Penny Yce
and Greg Pierce for their concern and support in my professional development.

I would like to thank Kate Hall for her invaluable help with the interviews. I
would also like to thank all of the staff and students at the Outcomes Evaluation Center
for their help with the data entry.

In addition, I must thank my family and friends for their support. In
particular, I would like to thank my parents, Susie and Geoffrey, for their
unconditional love and support. I would also like to thank Nancy, whose own
endeavors inspired me at times when I felt like giving up. I also would like to thank
Christina for her friendship and support throughout the process and her help with the
manuscript. Finally, I must thank David for encouraging me when I was down and

whose love, support and faith in me were invaluable.

iii

Table of Contents

LIST OF TABLES ........................................ vi
INTRODUCTION .......................................... 1
Level of Functioning Scales ............................... 3

Level of Functioning in Relation to Diagnosis and Symptomatology ...... 4

Prior Measures of Functional Impairment ....................... 7
Global Measures of Impairment ................... 7

Multidimensional Measures of Functional Impairment . . . . 15
Overlap between Functional Impairment, Symptom Distress & Diagnosis . 22

Informant Issues - ..................................... 25
Psychometric Evaluation of the FISCA ....................... 26
HYPOTHESES ........................................... 31
METHOD .............................................. 33
Sample ........................................... 33
Inpatients ..................................... 33

Outpatients .................................... 34

Measures .......................................... 36
Measures of Functional Impairment .................... 36

Measures of Symptomatology ........................ 38

Measures of Diagnoses ............................ 42

Length of Stay .................................. 43

Demographic Information ........................... 43
PROCEDURES ...................................... 44
Inpatients ..................................... 44

Outpatients .................................... 47

RESULTS .............................................. 49
Preliminary Analyses ................................... 49
Hypothesis 1A ...................................... 49
Hypotheses IB and 1C ................................. 50
Hypothesis 2A ...................................... 56

iv

Hypothesis 2B ....................................... 58

Hypothesis 2C ....................................... 61
Hypothesis 3 ........................................ 66
DISCUSSION ........................................... 67
Concurrent Validity of the FISCA .......................... 67
Discriminant Validity of the FISCA ........................ 72
Predictive Validity of the FISCA ........................... 75
Future Directions ..................................... 76
Methodological Problems ................................ 79
Conclusions ........................................ 79
Appendix A - FISCA Criteria .................................. 81
Appendix B - Time Line ..................................... 85
Appendix C - Parent or Guardian Informed Consent ................... 86
Appendix D - Minor Child Informed Assent Agreement ................. 88
LIST OF REFERENCES .................................... 89

 

LIST OF TABLES

Table l - Demographic Information for all Samples ....................
Table 2 - The Correlations Between the FISCA and the CIS and CGAS .......
Table 3 - Correlations Between Narrow-Band CBCL Scales and FISCA Scales
Table 4 - Correlations Between FISCA Scales and YSR scales .............
Table 5 - Correlations between Alcohol Drug Questionniare Scores and the FISCA
Table 6 - Stepwise Discriminant Analysis of Level of Care using the FISCA scales

Table 7 - Age, Home, Self-Harm, and Alcohol-Drug FISCA scales Discriminant
Function’s Ability to Predict Inpatients and Outpatients

Table 8 - Results of Analyses Using FISCA Scales to Distinguish Between .....
Diagnostic Groups

Table 9 - Ability of FISCA Scale Discriminant Functions to Classify Diagnostic
Group Membership

35

.53

55

57

59

.59

63

. 64

INTRODUCTION

The purpose of this study was to validate the Functional Impairment Scale for
Children and Adolescents (FISCA; Frank & Paul, 1995) by establishing its concurrent,
discriminant, and predictive validity. As such, the study assessed the degree of overlap
between functional impairment, as measured by parent report on the FISCA, and
patient symptomatology and diagnoses. The FISCA assesses children’s and
adolescents' functional impairment across a variety of areas and contexts. A major
interest of this study was to demonstrate that measures of functional impairment are
related but not identieal to measures of symptomatology and diagnosis. As such, I
hoped to Show that measures of functional impairment in children and adolescent are an
important if not essential addition to traditional assessment protocols that focus on
symptoms and diagnostic eategories in assessing psychopathology in children and
adolescents.

Child and adolescent inpatient treatment has undergone a profound
transformation in the past two decades. Twenty years ago, most inpatient facilities for
children and adolescents relied on long-term treatment (Nurcombe, 1989). Treatment
focused on a detailed exploration of the patient's psychic conﬂicts and their origin, and
an assessment of ego defenses and ego strengths. However, two major factors are

making this model of treatment increasingly “out of date”.

 

2

First, changes in the economy have led to increased costs of inpatient eare and a
greater reliance on third-party payers, with accompanying demands for briefer effective
treatments, shorter stays and reduction in costs. Second, the incidence of reported
eases of children and adolescents with severe psychopathology has increased
dramatically in the last two decades. Currently, more than twice as many children and
adolescamarebeingueatedmmsidenﬂalcmtemaswmpamdmalmMMOdecades
ago (Taube and Barret, 1985). As a result of both of these factors, inpatient facilities
have found themselves with increasingly limited resources to care for increasingly
larger numbers of children. Consequently, the criteria for inpatient admission has
become more stringent and the length of stay of patients has decreased.

These changes have necessitated signiﬁcant changes in both treatment goals and
in the nature and time frame of treatment procedures for children, adolescents and their
families. Whereas traditional, psychoanalytically-oriented, long-term treatment focused
on a total reconstruction of the personality (Nurcombe, 1989), the major goal in most
facilities today is to get the patient out of the hospital. Long-term treatment required
an assessment of such things as prior history, intrapsychic processes (defenses, ego-
strength, etc.) and perceptual styles, whereas short-term treatment requires an
assessment of clear and speciﬁc problems that make it impossible for the patient to
function in a non-restrictive environment.

Brief inpatient treatment should create the changes necessary for the child to
leave the hospital and continue treatment in a less restrictive and less expensive
environment. According to Nurcombe (1989), the central aim of treatment is

stabilization of those problems that are preventing the patient from being treated safely

3

outside of the hospital. Assessment conducted at admission to the hospital should be
brief and focused on identifying the pivotal problems. Assessment needs to look at
which symptoms, signs, impairment, behaviors, emotions, dispositions, or dysfunctions
in the current biopsychosocial pattern must change if the patient is to be treated at a less
restrictive level of care (Nurcombe, 1989). An obvious goal of assessment is to
identify a patient's current functional difﬁculties so that the targets of brief
hospitalization can be determined and addressed. When these problems have stabilized
and it is safe for the patient to be treated in a less expensive and intrusive setting, the
child is discharged from the hospital.

To accomplish these kinds of assessment goals, a standardized, well-developed
tool for measuring children's and adolescents' functional impairment in a variety of
areas is needed. Several measures of functional impairment in children and adolescents
have been developed; however these measures have several flaws. In the remainder of
this chapter, I critieally review the literature on existing tools for measuring children's
functional impairment and discuss the development of an instrument that is intended to
address a number of flaws found in current measures of children's functioning.
LEW

In the past decade, there has been a burgeoning interest in studying children and
adolescents' functional impairment. In DSM-IV, an entire axis, axis V, is designated
to rate functional impairment, and research has indicated that the diagnostic variable
most strongly associated with hospitalization is failure in adaptive functioning
(Mezzich, Evanczik, Mathias, & Coffman, 1984). The importance of assessing

impairment is also supported by research indicating that presence of criteria for making

4
a diagnosis does not necessarily imply need for treatment (Bird et al., 1990).
Researchers have also found that impairment criteria are sensitive predictors of
treatment outcome in clinieal trials (Gordon, Plutcky, Gordon, & Guerrra, 1988).

A level of functioning scale, when used properly, can provide a systematic ,
reliable, and efﬁcient form of reference for formulating and communicating how the
client is adapting to his or her environment, and what level of care they need. A level
of functioning seale is seen as a complement to, rather than a substitute for, a
diagnostic system or symptom scale. While diagnostic systems and symptom seales are
useful in identifying the nature of a client's disorder and indieative of distress
(symptoms), a level of functioning seale can be useful in describing the client's ability
to function at one point in time and over time (Newman, 1983). It ean be used to
ikmify to what extent the person has difﬁculty functioning in his or her environment at
any given point in time and what level of treatment is necessary so that the person can
ﬁrnctionatasatisfactorylevelinthefuture. Inadditionitcanbeusedtoassess
changes in patient functioning over time.

I l E 5 . . . l . l' . l l
The importance of the distinction between a patient's impairment in functioning and
his or her symptoms or diagnosis has a long history (Weissman et al., 1983). In both
theory and current assessment practices, level of functioning seales, diagnosis, and
level of symptomatology are all distinct entities, although there is some degree of
overlap among the three.
Level of symptomatology typically is assessed using problem or symptom

checklists (6.3., Achenbach & Edelbrock, 1983, Quay, 1977) that allow for

5

comprehensive ratings of problematic and behaviors as well as unpleasant subjective
experiences. Although no research has been conducted on the overlap of level of
symptonutology and functional impairment, it makes same to hypothesize that there
should be some relation between the two. A relation is expected because functional
impairment assesses disruption in a child’s ability to function as a result of disruptive
and symptomatic behavior. For example, depressive symptoms, such as difﬁculty with
concentration, psychomotor retardation, and anhedonia could affect an adolescent‘s or
child's intellectual achievement, resulting in poor functioning in school. The causality
of the relationship between functional impairment and symptoms may go in either
direction or in both directions. For instance, severe impairment ean result from high
levels of symptomatology, but high levels of symptomatology do not necessarily imply
severe impairment. A particular patient may have many symptoms, but these
symptoms may not interfere with the patient's ability to function in his or her
environment. Conversely, impairment in functioning could, but need not necessarily,
lead to symptoms. For instance, school failure could (but need not) lead to or predict
depression or antisocial behavior.

There is also a certain degree of overlap between symptom groups and
diagnosis. The overlap is evident in the criteria used for making a diagnosis.
Diagnosis of a disorder is based on several speciﬁc criteria that must be met. Typically
the criteria include a speciﬁed number of symptoms from a larger list, duration
requirements for the symptoms, and various exclusion criteria that state that the
symptoms are not due to some other condition. In some instances, certain symptom

groups are the only or the major deﬁning features of the diagnosis, so that a great deal

 

6
of overlap between symptomatology and diagnoses is to be expected. For instance,
certain symptom groups or behavior clusters associated with delinquency might relate
strongly to DSM-IV diagnosis of conduct disorder. There may also be a correlation
between symptom groups and diagnoses, even when the symptom seale contains only
one or two items required for the diagnosis and also contains nonrelated items. For
instance, thereshouldbeamoderatetohigh correlation betweenadiagnosisof
depression and somatic symptoms, which include not only items such as "overtired' (an
item used as a deﬁning criteria for a diagnosis of depression), but also items not used in
deﬁning affective disorders, such as nausea, pains and headaches. Researchers have
reported substantial overlap between empirically derived symptom syndromes and
DSM-III diagnosis (Edelbrock and Costello, 1988; Weinstein, Noam, Grimes, Stone,
& Schwab—Stone, 1990). These studies found a signiﬁcant overlap when the symptoms
scales were part of the criteria for diagnosis and also when the symptom groups
contained only one or two items that were part of the criteria for diagnosis but
contained nonrclated items as well. Not too surprisingly, the overlap was greatest
when the symptom groups were part of the criteria for diagnosis.

There should also be an overlap between diagnoses and measures of level of
impairment, although far less work has been done in this area. Prior to DSM-IV this
overlap might have been more minimal because in order to receive a diagnosis the
child or adolescent did not have to be functionally impaired. In fact, many children
who met diagnostic criteria, according to DSM-III-R, were not impaired in their ability
to function in their environment (Bird et al., 1990). Currently, in order to receive

many of the DSM-IV diagnoses the child or adolescent must also be impaired. For

7
instance, in order to receive a diagnosis of depression the symptoms must cause
clinically signiﬁeant distress or impairment in social, occupational or other important
areas of functioning. Impairment, like symptomatology, now is embedded in the
criteria for diagnosis; therefore it makes sense that impairment will overlap with
diagnosis much as symptomatology does. However, there are no studies assessing the
extenttowhichthisistrue. Inthisstudy, Iwillassesstherelationshipbetwoenanew
measure of functional impairment and level of symptomatology and diagnosis.
W

GlobalMeasuresnﬂmpairment. An important hypothesis in this work is that
the new instrument to be presented here improves in a number of ways on earlier tools
used to measure functional impairment in children and adolescents. One advantage is
that the FISCA does not depend on clinieal judgments based on ambiguous data
sources.

The ﬁrst measure of functional impairment in children and adolescents was the
Global Assessment Scale for Children. The GAS-Children rates functioning on a scale
of l to 100. A descriptive paragraph explains the criteria for each 10-point range. For
example, a score from 1—10 is given if the patient ”needs constant supervision (24-hr
care) due to severely destructive behavior or gross impairment in reality testing,
communication, cognition, affect, object relations, or personal hygiene” (Rotham,
1976). Inter-rater reliabilities were .56 for inpatients and .84 for outpatients
(Sorenson, 1982). However, the primary rater provided the secondary rater with
information about the patient, which could have inﬂated these reliability coefﬁcients.

As would be expected, the mean score for inpatients was higher for outpatients

(51.8) than for inpatients (32.5). However, Sorenson et a1. (1982) looked at how
differentratersratedthestandardvignettesgivenintrainingoftheGASandfound that
inpatient raters rated the standard vignettes as more impaired than the outpatient raters
did. This bias easily could have led inpatient subjects to be rated as more impaired. A
structured, objective scoring system would of course eliminate this bias.

The validity of the GASC has not been supported in the literature. Sorenson,
Hargreaves, & Friedlander (1982), reported nonsigniﬁcant correlations between the
GASC and the total problem score of the Conners Parent Questionnaire (Conners,
1989) and with total raw score of the Achenbach Child Behavior Checklist (Achenbach
& Edelbrock, 1987); and Keraus (1981) found generally insigniﬁcant correlations
between the GASC and summary scores on the Behavior Problems Checklist (Quay,
1977). Interpretation of these studies is somewhat problematic since symptoms and
impairment are not equivalent; however a certain degree of overlap ought to be evident
between the two. No other validity studies have been published. A review of recent
studies assessing children's impairment, suggest that the GASC is no longer being
used. The most widely used measure of functional impairment in children and
adolescents is the Children's Global Assessment Scale (Shaffer, Gould, Brasic,
Ambrosini, Fisher, Bird, & Aluwahlia, 1983). It is a measure of global impairment in
children's functioning during a speciﬁed time period. It is based on the Global
Assessment Scale, (Endicott, Spitzer, Fleiss, & Cohen, 1976), a measure of overall
functional impairment for adults. The Children's Global Assessment Seale (CGAS) is
designed to reflect the lowest level of functioning for a child or adolescent during a

speciﬁed time period. Its values range from 1, representing the most functionally

9

impaired child, to 100, representing the highest functioning child. Scores above 70
indicate normal functioning. The scale contains behaviorally oriented descriptors at
each anchor point that depict behaviors and life situations applieable to children 4
through 16 years of age. Studies of the psychometric properties of the CGAS have
used case histories (Shaffer et al., 1983; Steihausen, 1987), video tapes (Bird, Canino,
Rubic-Stipee, & Ribera, 1987), in-person interviews (Apter, Orvashel, Laseg, Moses,
& Tyano, 1989), and ratings by clinical staff who were directly involved in the clinical
care of the patients studied. The intraclass correlation coefﬁcient of different raters in
the different studies was high, ranging from .72 to .93. Bird et al., reported
satisfactory test-retest stability (.85) over a 19 day interval. The CGAS was also found
to discriminate between outpatients (65.4, SD =14.8) and inpatients (46.0, SD =19)
(Shaffer et al., 1983). This difference was signiﬁcant at the .001 level. The
concurrent validity of the CGAS was studied in relation to measures of symptom
distress and child competence. The correlation between the CGAS and the Conners
Abbreviated Seale, which measures symptom distress, was -.25 (p > .05, df= 17) in an
outpatient sample (Shaffer et al., 1983). This correlation suggests that the Conners
index and the CGAS tap somewhat overlapping but different domains of assessment.

The CGAS was also studied in relation to the total CBCL score (Achenbach and
Edlebrock, 1981), but ﬁndings on the relationship between these two measures have
been inconsistent across studies. A highly structured study using psychiatrists and
psychologists ratings of randomly sampled children in a Puerto Rican community,
found a Pearson correlation between the CGAS score and the total behavior problem

score on the CBCL of -.65 (Bird et al., 1990). However, when data were collected as

10

part of a clinical assessment, using psychiatrist ratings and milieu staff ratings in an
inpatient hospital, the investigators found a nonsigniﬁcant correlation of .-05 for the
attending psychiatrist's ratings on the CGAS and total CBCL score, and -.11 for milieu
staff ratings on the CGAS and total CBCL score (Green, Shirk, Hanze, & Wanstrath,
1994). The differences between these two studies indicate that clinical judgement in a
controlled research setting and in a clinieal setting may lead to different results.

Findings on the relationship between the CGAS and the Social Competence
scale of the CBCL are also inconsistent. Bird et a1. (1987) found a signiﬁcant
correlation (.58), whereas Apter et a1. (1989), Vandvik (1990), and Green et al. (1994)
found nonsigniﬁcant correlations.

Green et al., (1994) found that the CGAS correlated signiﬁcantly with a number
of indicators of individual functioning. The CGAS correlated with Activity
Competence (.26); Full Seale IQ (.43); and Social Relatedness (.33 for psychiatrist
ratings, and .39 for milieu staff ratings). In addition, Vandvik (1990) reported a
signiﬁcant correlation between the CGAS (r=-.80, p < .010) and the total score from
a structured diagnostic interview, the Child Assessment Schedule (Hodges, Kline,
Stern, Cytryn, & McKnew, 1982).

While the CGAS appears to have satisfactory reliability, the concurrent validity
of the measure varies across studies. Also, the CGAS has been tested only with
professionals. One major problem of this instrument is that it would be difﬁcult to
implement in a variety of settings because the ability of para-professionals, who often
are responsible for intake prowdures, to use the scale accurately has not been shown

andtherearenotraining manuals.

11

The Health and Sickness Rating Scale for Children (HSRS-C; Liebowitz,
Rembar, Kornberg, Frankel, & Kruger, 1988) is another impairment measure used to
assess children’s psychiatric disorder. It was modeled after the Health-sickness Rating
Scale for adults (HSRS; Luborsky, 1962). It was designed for use with children 6 to
11 years old. The scale ranges from O to 100. The anchored scale points range from
9, a point of extreme disturbance, to 94 which indicates superior adaptive,
interpersonal, and academic functioning. The endpoints of 0 and 100 are reserved for
theoretical extremes of disturbance and adaptation. Criteria are spelled out for each
anchor point. Liebowitz et a1. (1988) examined the interrater reliability of clinician's
individual ratings and reported a intraclass correlation coefﬁcient of .73. Liebowitz et
al. (1988) also studied the validity of this measure. Discriminant validity was assessed
by comparing HSRS-C scores for inpatients and outpatients. The mean rating HSRS-C
rating for the outpatients was 59.1 and 38.8 for the inpatients. Concurrent validity was
assessed with the Child Behavior Checklist (Achenbach and Edelbrock, 1983) and the
Child Assessment Schedule (CAS) (Hodges et al., 1982). Negative correlations of -
0.71 and -0.75 with the total CBCL behavior problem score and the externalizing scale,
respectively, were reported. Only a modest correlation of 0.47 was found with the
internalizing scale, possibly indicating that internalizing patients are less likely to be
functioning poorly than externalizing patients. A signiﬁcant negative correlation of -
0.48 was found with the CAS.

Although Liebowitz et a1. (1988) reported that the HSRS-C has good reliability
and validity, there were several ﬂaws in this study. The study was conducted using

trained clinicians as raters and the authors assume that only highly trained clinicians

12

will use this seale. Therefore, the ability of less trained clinicians or paraprofessionals
toreliablyusethismeasurehasnotbcenstudied. Inaddition, thecaseswereratcd
from written protocols of actual case histories or comprehensive reports, which may
havecontainedcuesabouttlnelevelofcarethepatientswerereceiving.

A fourth measure of functional impairment is the Global Level of Functioning
Selle (GLOF; Hodges, Bickrnan, & Kurtz, 1991), a modiﬁed version of the Child
Global Assessment Scale (Bird et al., 1987). The GLOF is used to generate an overall
severity of impairment scores, ranging from 0 to 100. The GLOF has a comprehensive
training package which includes detailed scoring instructions and examples. Raters are
instructed to consider the child's functioning in four major areas: home/ family, school,
peer/social relationships, and community. Hodges et al. (1991), looked at the
reliability of 13 raters who were not trained in working clirnically with children; in a
second study she assessed the reliability of ratings from 20 ﬁrst year graduate students
in clinical psychology was assessed. In both studies, the raters scored twenty vignettes
which consisted of a brief description of the family constellation and currernt
circumstances and a summary of the child's responses on the CAS (Hodges, 1989).
They reported interclass correlations for two studies, .79 and .81 respectively. There
have been no studies of the validity of this measure.

All of the scales discussed so far require clinical judgement. Bird, Shaffer,
Fisher, Gould, Staghezza, & Hoven (1993) saw the need for a measure of functional
impairment that did not require clinical judgement, and hence, developed the
Columbia Impairment Scale. A respondent (e. g. , parent or caregiver) is asked 13

questions used to assess four areas of functioning: interpersonal relations, broad

13

psychopathological domains, functioning at work or school, and use of leisure time.
Each question is rated by the respondent on a ﬁve point continuum of 04; 0, no
problem; 1-3 some problem; and 4, a very big problem. Factor analysis and the
assessment of the internal consistency of this scale using a pilot sample, led Bird et al.
(1993) to conclude that the scale assesses a single factor. Although the authors wanted
toassessfunctionalimpairmentinseveralareas, itappearsthatthisisameasureof
global functioning.

ThereiscurrentlyonlyonestudyontheCISanditusespilotdatafroma
community sample (Bird et al., 1993). Bird et al. (1993) had both parents and children
answer the questions. The psychometric properties of the parent instrument appear to
be better than those of the child and adolescent instrument. Bird et al. (1993) report
good test-retest reliability for parent-report (Cronbach's alpha=0.89) and signiﬁcantly
lower reliability for child-report (Cronbach's alpha= .63). They also report that the
Columbia Impairment Scale correlates signiﬁcantly with a therapist's ratings, r=-0.56.
(The negative correlations relate to the fact that the two scales are scored in opposite
directions). The CIS also correlates with other indicators of psychological dysfunction.
Grades in school correlated with parent C18 (.45) and with child C18 (.30). The parent
CIS correlated with whether the child had been expelled from school (.32) but the
ﬁndings were insigniﬁcant for child C18 (0.08). Child's adaptive competence
correlated with parent C18 (-0.71) and with child C18 (-0.37). The authors did not
analyze the child and adolescent respondents separately and therefore the impact of age
on CIS reports provided by children is not known.

The Columbia Impairment Scale does not require clinical judgement, which

 

14

allows it to be more easily used by a wide array of individuals; however because the
scale does not provide the respondent with clear behavioral or symptomatic criteria it is
relatively susceptible to the subjective judgment of the respondent. This could be
problematic in the assessment of a clinical sample because some studies have concluded
that maternal perceptions of child adjustment and functioning are related to maternal
psychopathology, marital discord, expectations for child behavior, self-esteem,
stressors and social support (Ferguson & Horwood, 1987; Forehand, Lautenschlager,
Gaust, & Graziano, 1986; Mash & Johnston, 1983; Moretti, Fine, Halye & Marriage,
1985). The more open-ended a scale, and the less deﬁned the criteria for making
ratings, the more likely that these confounds will occur.

Axis V in DSM-IV is a ﬁfth measure of impairment (American Psychiatric
Association, 1993). This axis uses the Global Assessment of Functioning (GAF) Scale
to rate both children and adult's overall functioning. It is to be rated with respect only
to psychological, social, and occupational functioning. The GAF is rated on a scale
from 1, indicating the lowest level of functioning to 100, indicating the highest level of
functioning. Each ten point interval has a symptomatically oriented description. The
GAF was derived from the GAS and the CGAS. There are no published reliability or
validity data for this measure. It is supposed to be rated by a clinician. This could
lead to difﬁculties with the reliability of the rating, since it is difﬁcult for most
clinicians to avoid confounding diagnostic speculations with assessments of a patient's
functioning.

Axis V was ﬁrst included in DSM-III (APA, 1980). This was a seven point

scale, in which a clinician was asked to rate the client's highest level of adaptive

 

15
functioning during the past year. Rey, Plapp, Stewart, Richards, Bashir (1987), found
that reliability for Axis V in DSM-III (1980) was similar for younger (.63) and older
boys (.56), but reliability for younger girls (0.36) was lower than for older girls
(0.51). A validity study by Rey, Stewart, Plapp, Bashir, & Richards (1988) found
that, for adolescents, Axis V correlated more highly with premorbid functioning (0.76)
than with present social competence (-O.46).

All of the measures discussed thus far are global measures of
impairment. Global measures give a description of the child's overall level Of
functioning, but do not discretely measure the child's functioning in a variety of areas.
Multidimensional measures have been developed to provide more elaborate
information on children's functioning.

The Childrern's Impairment Scale, (Sorenson, 1982) is one such measure. It
consists of global ratings in four areas: Developmental Status (overall maturity of a
child's physical, emotional, and intellectual capabilities), School Adjustment (child's
academic performance and behavior at school), Interpersonal Relations (child's ability
to relate to others), and Current Living Environment (the ability of the living
environment to meet the child's needs), with no overall rating of function. Each of
these subscales is scored by assigning a number from one to ﬁve, where level ﬁve
represents severe impairment and level one represents no impairment. A descriptive
paragraph explains each level of each scale. The inter-rater reliability for the
Developmental Status was .69 for outpatients and .40 for inpatients; .69 for outpatient

school adjustment, .38 for inpatient school adjustment; .69 for outpatients interpersonal

16
relations, .22 for inpatient interpersonal relations ; .81 for outpatient living
environment and .39 for inpatient living environment (Sorenson, 1982). All of the
scales had higher reliability coefﬁcients for outpatients than for inpatients. This may
reﬂect the greater range of impairment among outpatiernts compared to inpatients, or
it may reﬂect a true superiority in this scale's performance in outpatient settings. All
of the scales except for the Developmental Status scale were able to discriminate
between inpatients and outpatients. However, different therapists rated inpatients than
those who rated outpatients. Similar to the GAS, Sorenson et al. (1982) found that
inpatient raters rated the training vignettes as more impaired than outpatient raters. The
different areas of impairment were modestly intercorrelated in adolescents, but among
young children, the correlations among the subscales were high, ranging from .60 to
.80. The Children’s Impairment Scale was not studied in relation to other measures of
child psychopatlnology, so the concurrent validity of this measure has not been
established.

The Progress Evaluation Scales for Children (ages 6 to 12), arnd the Progress
Evaluation Scale for Adolescents (ages 13 to 17) (Ihilevich & Gleser, 1982) are also
multidimensional measures of functional impairment. The PES rates functioning in a
variety of domains including: 1)Family interaction (dependence—independence-
interdependence in one's relationship with other family members); 2) occupation
(ability to function in school, job, or homemaking role); 3) getting along with others
(socialization); 4) feelings and mood (the level of affective modulation); 5) use of free
time (ability to participate in and create resources for play and enjoyment); 6) problems

(tine coping capacity the person can bring to bear on his or her daily problems); and 7)

17 :

attitude toward self (self-esteem). Each scale consists of ﬁve levels rated from 1 to 5,
from the most pathological to the healthiest levels of functioning observed in the
community. The scales can be completed by parents, therapists, and the child or
adolescent. A study using therapists as informants found that the scales were able to
differentiate between outpatiernts and patients in partial hospitalization, with children in
partial howitaliaation scoring lower (i.e., more impaired) on all of the PES scales.
The adolescent self-report version of the PBS was compared to the self-report
Adolescent Life Assessment Checklist (ALAC; Gleser et al., 1977), that measures
affective distress, unproductivity, sociopathy, peer alienation, somatic complaints, and
tolerance of intimacy. All but the last scale of the ALAC, are keyed so that a higher
score indicates more pathology. Resulting correlations indicated a high degree of
concurrent validity between the two instruments. Family Interaction correlated .26
with Tolerance of Intimacy. Poor Occupation (school adjustment) related signiﬁcantly
with Unproductivity (-0.24), Sociopathy (-0.28), Per Alienation (-0.32), and Somatic
Complaints (-0.51). Difﬁculty in Getting Along with Others was associated with
Sociopathy (-0.20), Peer Alienation (—0.26), and poor Tolerance of Intimacy (0.19).
Poor Use for Free Time was also associated with Peer Alienation (-0.31) and poor
Tolerance of Intimacy (0.30). Both Feelings and Mood and Problems correlated
signiﬁcantly with ﬁve of the six scales.

Although, the PBS is noteworthy in its attention to a wide variety of functional
areas, it has a number of drawbacks. There is an unduly large inter-rater variance on
some scales. The estimated variance due to average differences between ratings of

therapists on any one person ranged from .18 to .52 in the outpatient sample and .18 to

18
.50 for the day hospital. In addition, inter-rater reliability was only assessed on a very
small sample of children (N =14). Ihilevich and Gleser (1982) hypothesized that the
source of the difﬁculty lay in the fact that children's behavior is often difﬁcult to
interpret and parental reports are frequently inconsistent. To improve the quality of the
data, they suggested that clinician ratings of children on the PES should be made after
two or three intake sessions, instead of the previously used one hour diagnostic
interview. While this would help establish rapport and allow the therapists a greater
period of time in which to judge the child, in an inpatient setting, it is too lengthy and
laber intensive. It does not brieﬂy assess the pivotal problem, so that a treatment plan
can be implemented immediately upon admission to the facility. Another problem with
the PBS, is that it is not sufﬁciently behaviorally anchored and items do not have
objective referents and hence are vulnerable to subjective interpretations.

In the adult literature, alternative measures have been developed that attempt to
identify impairment in multiple areas. One such measure is the North Carolina
Functional Assessment Scale (NCFAS; North Carolina Department of Human
Resources, 1989). Hodges (1991) used the NCFAS as a model to develop the Child
and Adolescent Functional Assessment Scale (CAFAS), a multidimensional measure of
functional impairment in children. In developing the CAFAS, extensive modiﬁcations
were made to the NCFAS so that it would be more applicable for children and
adolescents. The CAFAS is a scoring grid and not a questionnaire. Hodges consulted
with 40 psychologists and psychiatrists regarding the face validity of the measure.
These consultants were able to provide a wide array of perspectives including, child

psychopathology, normal development, and the special needs of Hispanic and African-

19

American children. However, the CAFAS is not based on any particular theory or
model of psychopathology and ratings are not intended to reﬂect the etiology/causes or
the dynamics underlying the youth's problems or dysfunctions. Instead the CAFAS
mostly measures the degree of impairment in a youth's or the child's current
functioning regardless of history, causes, or prognosis of a mental health disorder.

The CAFAS yields a total score as well as scores for subscales, consisting of
Role Performance (in home/work, school, and the community), Thinking, Behavior
Toward Others, Moods/Emotions, Self-Harm and Substance Use. For each subscale, a
rater determines the most severe level of dysfunction within a speciﬁed period of time
(usually 3 months). Four different levels of impairment can be assigned. These four
levels are: 0 for Average (minimal or no impairment); 10 or Mild (signiﬁcant problems
and distress); 20 for Moderate (persistent disruption or incapacitation); and 30 for
Severe (severe disruption or incapacitation). For each level, on every scale there are
items with specifying criteria for that level. To score each subscale, the rater reviews
the items in the Severe category ﬁrst. If any item describes the child's functioning, the
rater assigns a score of Severe or 30 to that subscale. Ifnone ofthe Severe items
describe the child, the rater precedes to the Moderate category, and progresses in this
manner until an item that describes the child's functioning is located. The rater uses
information obtained from a semi-structured interview with the child's parent to decide
on the child's level of impairment. Raters can use additional information from the
Child Assessment Schedule (CAS; Hodges, Cools, & McKnew, 1989) and information
from case records to rate the level of impairment.

The CAFAS scale originally was used as a guide for rating functional

20

impairment on the bases of information obtained from the Child Assessment Schedule,
a structured diagnostic interview (Hodges et al., 1989). Two studies assessing
interrater—reliability used a brief description of the family constellation and current
circumstances, and a summary of both mother and child responses on a structured
diagnostic interview (i.e., Child assessment schedule; Hodges et al., 1989) to obtain
CAFAS scores. The ﬁrst study used 13 raters, six of these raters had college degrees
orless, sixhadmastersdegrees, andonehadadoctoraldegree. Thesecond studyused
20 raters who were enrolled at Masters program in a Clinical Psychology. Raters for
both studies were trained with a manual. Raters for the second study were also given
supplemental guidelines for scoring the CAFAS. The raters rated each subscale and
added each level of impairment to obtain a total score. The results for the total
CAFAS score (Study 1: r=.82; Study 2: r=.81) and for Role Performance (Study 1:
r=.74; Study 2: r=.74) and Behavior Towards Others/Self ( Study 1: r=.77; Study 2:
r= .78) were satisfactory in both studies. The results for the Moods/ Emotions subscale
were unsatisfactory in the ﬁrst study (r= .44), but were slightly better in the second
study (r= .69). The results for the Thinking subscale were somewhat unsatisfactory in
the ﬁrst study (r=.64) and very unsatisfactory in the second study (r=.3l). The
Substance Use subscale was assessed in the second study only (r=.86). Pearson r
correlations calculated between the GLOF and the CAFAS total score were signiﬁcant
(r=-.84 for Study 1 and -.83 for Study 2). The negative correlations were expected
because high impairment is noted by a low score on the GLOF and a high score on the
CAFAS.

The CAFAS Scale uses clearer and more objective referents than its

21

predecessors. However, while the CAFAS measures a multidimensional level of
functional impairment, it still uses raters judgements to obtain scores, and these
judgements are still subject to personal biases (as noted by difﬁculties obtaining
adequate reliability on several scales). A new semi-structured interview designed to
accompany the CAFAS scale, allows the rater to probe for more information.
Reliability of this instrument is currently being assessed (Hodges, personal
communication). However, the extent to which each individual rater probes for
information could affect the score that the youth is given. No data on this hypothesis is
available, but it is a potential ﬂaw in this instrument.

Frank & Paul (1995) recently developed an alternative method of assessing
functional impairment that in large part was based on the CAFAS but relied solely on
parent report rather than clinicians judgements based on interviews with patients. This
method uses an objective questionnaire to assess how impaired a child is in the same 8
areas measured by the CAFAS scale. These areas are School, Home,
Community/Legal, Thinking, Being with Others, Moods and Emotions, Self-Harm,
and Alcohol and Drug Use. For each area of the FISCA scale, there is a corresponding
section in the FISCA questionnaire. The purpose of this study is to examine the
psychometric properties of this questionnaire, known as the Functional Impairment
Scale for Children and Adolescents (FISCA; Frank & Paul, 1995).

Briefly, the development of the scale entailed taking each scale on the CAFAS
(e.g., school) and combining similar criteria (e.g., grade average is lower than 'C" and
failing at least half of academic courses). The authors also deleted criteria that used

prior history or mental illness to rate the child as more impaired (e.g., criteria that

2 2
rated the child as more impaired if they had received an attentional disorder diagnosis).
Criteria on the CAFAS that were overly detailed or redundant were simpliﬁed (e. g. ,
behavior causes removal from regular school (or impending threat of removal) due to
potential harm to others related to aggressive behavior or threat of aggressive behavior”
is now "behavior is so out of control that the child is practically unmanageable in the
classroom"). Criteria on the CAFAS scale that were ambiguous were clariﬁed. For
example, “not in schoOl because of impairment” was stated as "not in school because of
school refusal or school phobia”.

For each of the FISCA criteria Frank and Paul asked an objective question to
see if the child met the criteria. For example, for criteria 04 (chronic skipping or
truancy resulted in punitive actions or poor academic performance) parents were asked
how many times their child had skipped school and if at all, whether skipping resulted

in lost course credit, poor grades, complaints from school ofﬁcials or suspension.

 

The remainder of this chapter discusses issues inherent in assessing the

psychometric properties of the Functional Impairment Scale for Children and
Adolescents (FISCA). The FISCA and widely used measures of symptom distress such
as the Child Behavior Checklist (CBCL; Achenbach & Edelbroch, 1983) and the Youth
Self-Report (YSR; Achenbach & Edelbroch, 1987) measure different dimensions of
child and adolescent psychopathology. These measures in certain respects are
structurally similar. The FISCA measures speciﬁc areas of impairment as well as a

total level of impairment; the CBCL and YSR measure groups of symptoms or

23

syndromes and a total level of symptomatology.

The CBCL and YSR measure eight ”core syndromes“ or symptom groups.
These core syndromes are labeled Withdrawn, Somatic Complaints,
Anxious/Depressed, Social Problems, Thought Problems, Attention Problems,
Delinquent Behavior, and Aggressive Behavior. In addition, the narrow-band
Withdrawn, Somatic Complaints, and Anxious/Depressed scales together make up an
Internalizing factor and the narrow-band Delinquent Behavior and Aggressive Behavior
scales together constitute an Externalizing factor. Most (118 of the 120) individual
items are combined to make a total problem score.

The CBCL syndromes were common to nearly all age and sex groups and to
both the parent report on the CBCL and child report on the YSR (Achenbach, 1991). In
addition to these scales, the CBCL and YSR include scales to measure social
competence. The CBCL measures social competence in three areas: activities, social
andschool. Thesethreeareascanalsobecombinedtoformatotalcompetence score.
The YSR only measures social competence in activities and social settings.

A certain degree of overlap is expected between the FISCA and measures of
symptomatology such as the YSR and the CBCL because the FISCA includes some
symptoms/behaviors as part of the criteria for measuring impairment. Areas of the
FISCA that are in large part deﬁned by the presence of symptoms (e.g., Thinking)
should overlap quite a bit with scales on the CBCL and YSR that measure that
symptom group (i.e. , Thought Problems). However, while a reasonable degree of
overlap (i.e., a signiﬁcant correlation) is expected, a certain degree of independence

(i.e., correlation should not be more than .70) between the FISCA and the CBCL and

24

YSR is also anticipated. Other measures of Functional Impairment have correlated
with symptom checklists from .26 to .75. Similar results are expected with the FISCA
and measures of symptomatology. Symptom checklists require only that the patient
have certain symptoms; while in order to be rated as impaired (in Thinking as well in
other symptom areas such as Moods and Emotions) by the FISCA, the patient must not
only have the presence of symptoms but the presence of these symptoms must also be
linked to impaired functioning.

The structure of the FISCA also is similar in some important ways to DSM-IV.
Scoring of impairment is based on both symptoms and evidence of functional
impairment in the various areas measured by the FISCA. This structure is similar to
the DSM-IV in that groups of DSM—IV diagnoses are indicated by a combination of
symptoms and functional impairment. Considering these similarities, a signiﬁcant
degree of overlap between the FISCA scores and diagnosis is to be expected as well.
However, there are also differences between diagnostic categories and what is
measured by the FISCA as a result of difference in emphasis. The FISCA includes
symptoms, yet the main focus is on levels of functional impairment in speciﬁc areas; in
contrast, the main focus of diagnoses are particular clusters of symptoms not areas of
impairment. A child who is rated as impaired in only one area in the FISCA could
receive several diagnoses. For instance, a child who meets criteria on the FISCA for
severe impairment in school could meet criteria for several diagnoses that are often
associated with and in part identiﬁed with school impairment, including, depression,
mania, conduct disorder, and attention deﬁcit disorder. The diagnosis would depend

on what other symptoms accompanied the school impairment. Because of these

25

differences, a certain degree of independence between the FISCA and diagnoses should
be expected as well.
Informantlssusa

When studying the overlap between functional impairment and symptoms or
diagnoses, it is very important to consider from whom the information is being
obtained. The question of which informant provides the mom valid information as to
childhood behavior disorders, has been heavily researched. Consistently low
correlations between children, parents, teachers, peers, and mental health professionals
have been reported (Achenbach, McConaugh, 8s Howell, 1987; Verhulst & Van der
Ende, 1991; Kolko & Kazdin, 1993). These discrepancies have been found with
symptom scales (Moretti, Fine, Haley, & Marriage, 1985; Achenbach, McConaugh, &
Howell, 1987; Verhulst & Van der Ende, 1992; Kolko & Kazdin, 1993) and with
diagnostic interviews (Edelbrock, Costello, Dulcan, Conover, & Kalas, 1986; Hodges,
Gordon, & Lennon, 1990). These differences in reporting may be because one or the
other informant is a poor reporter or it may be that different informants contribute
different, but in each case, valid information. Parents seem more capable of reporting
behavioral manifestations of emotions, such as overt expressions of depressed affect or
aggressive behavior, but have more difﬁculty reporting on their children's internal
feelings. Hence higher parent-child agreement has been reported for externalizing than
internalizing symptoms (Edelbrock, Costello, Dulcan, Conover, & Kalas, 1986).
Several studies have indicated that children are better informants regarding their
subjective, internal symptoms, while adults are more likely to over report behavioral

problems (Angold et al., 1987; Ivens & Rehms, 1988; and Jensen et al., 1989).

2 6

While parents appear less sensitive to internalizing symptoms in their children,
their reports are highly speciﬁc (i.e., when they do identify depression, they are
usually correct) (Angold et al., 1987). In addition, the differences in parent and child
reponofdepressimdonotmeanmatchﬂdmnamnecessaﬁlybeuermportemofmeir
depression. While research has shown that children are able to report their emotions
accmately and consistently (Reynolds & Graves, 1989), this does not mean parents’
report of child's depression is not important. Parent and child report of depression
have been shown to relate to different problems (Kazdin, 1990). For instance, child
self-report measures of depression correlate with suicidal attempt and ideation,
hopelessness, low self-esteem, negative attributional style, and child—rearing practices
of the parent such as abuse (Haley, Fine, Mage, Moretti, & Freeman, 1985; Kazdin,
French, Unis, Esveldt—Dawson, & Sherick, 1983; J. Kazdin, Moser, Colbus, & Bell,
1985; and Sacco & Graves, 1984). Parental report of their children's depression
correlates with diminished social interaction patterns on the part of the child and overt
signs of expressive affect (Kazdin, 1985). Thus, it seems that both child and parent
reports are valid measures of childhood depression. Overall, the research on child-
parent differences indicates that regardless of what measure is used there will be a
discrepancy between the child's or adolescent's self-report and parent's reports.
PsychomtrieeyahrationnﬂheEISCA

This study attempts to provide some initial evidence of the FISCA's validity
using parent and child report data that primarily but not exclusively were collected in
an inpatient setting. This study focused on two types of validity including: construct

yalidity, which assesses the extent to which a test measures a psychological construct or

27

trait, and criterinnzrelatedxalidity, which measures the extent to which an instrument
measures or predicts some behavior as checked against an independent criterion
(Sattler, 1988).

Criterion-validity consists of three subtypes of validity: l)concurrent validity,
deﬁned by the strength of the relationship between the instrument and an alternative
measure of a similar construct obtained at approximately the same time; 2) discriminant
validity, assessed by the ability of the instrument to distinguish between theoretically
predicted groups; and 3) predictive validity, measured by the ability of the measure to
predict scores on another measure taken at a later point in time (Cronbach & Meehl,
1967). Since the FISCA will be used as an assessment tool, it is important to assess if
it is measuring what it is supposed to measure and the appropriateness with which
inferences can be made on the basis of the FISCA scores; i.e. the construct validity.
Because all three types of criterion-related validity can be used to demonstrate construct
validity ( Spitzer, Endicott,& Robbins, 1975), this study assesses the criterion-related
validity of the FISCA.

ConcurrenLYalidity: The FISCA was examined in relation to measures of
symptom distress to establish its concurrent validity. In order to assess the degree of
overlap between functional impairment and symptoms, the total FISCA score was
examined in relation to the total score on the Child Behavior Checklist (CBCL) and on
the Youth Self-Report (YSR). The FISCA total score was hypothesized to correlate
signiﬁcantly with the CBCL and the YSR and with youth report of drug and alcohol
use on a substance abuse questionnaire. . The literature indicates that child-parent

report discrepancies may result in slightly lower correlations between the FISCA

28

(parent-report) and the YSR (child-report) than between the FISCA and the CBCL
(parent-report). Each individual FISCA scale was hypothesized to correlate
signiﬁcantly with measures of symptomatology that assess conceptually similar problem
areas. While, the FISCA and symptomatology measures were hypothesized to be
signiﬁcantly related, correlations are expected to be moderate at best (i.e., around .35
to .40), since the FISCA and the symptomatology measures are not expected to be
identical.

In particular (1) the Moods and Emotions scale on the FISCA should relate to
the Withdrawn, Somatic Complaints, and Anxious/Depressed narrow—band scales; and
to the broad-band Internalizing Scale on both the CBCL and the YSR; (2) the Being
with Others scale should relate to the Delinquent Behavior, Aggressive Behavior,
Social Problems, and Extemalizing scales on the CBCL and the YSR; (3) the Thinking
scale of the FISCA should be related to the Thought Problems scale on the CBCL and
YSR; (4) the Alcohol and Drug Use scale of the FISCA should correlate with the
Adolescent Alcohol Involvement Scale (Mayer & Filstead, 1979) and the Michigan
Drug Use Questionnaire (Zucker, Noll, & Fitzgerald, unpublished instrument); (5) the
School scale of the FISCA should correlate with the Attention Problems, Aggressive
Behavior, and Delinquent Behavior narrow-band scales, as well as the Extemalizing
scales on the YSR and the CBC; and (6) the Home scale and the Community/Legal
scale of the FISCA should correlate with the Aggressive Behavior, Delinquent
Behavior, and the Exan scale of the CBCL and the YSR.

To assess concurrent validity of the FISCA, the total FISCA score was

examined in relation to two other measures of functional impairment: the Columbia

29

Impairment Scale (CIS; Bird et al., 1993) and the Children's Global Assessment Scale
(CGAS; Shaffer et al., 1983) . The total FISCA score was hypothesized to be related
toboththeCIS scoreandtheCGAS.

Ih'scﬁminantyalidity:1hesecondtypeofvaﬁditydratwasevaluatedis
discriminant validity. A comparison of ratings of impairment for inpatients and
outpatients has been consistently used to provide information pertaining to the
development of impairment measures (Shaffer et al., 1982; Bird, 1993; Sorenson et al.,
1982; Hodges et al., 1990). To evaluate the discriminant validity of the FISCA, mean
scores for inpatients and outpatients will be compared. It was hypothesized that
inpatients will have higher levels of impairment than outpatients.

Psychiatric diagnoses generated from the Schedule for Affective Disorders and
Schizophrenia for School-Age Children (K-SADS; Puig-Antich & Ryan, 1986), will
also be used to test the discriminating power of the FISCA. Each of the individual
FISCA scales were examined in relation to the scale’s ability to discriminate between
diagnostic groups that are conceptually similar to the FISCA scale in question.

The Thinking Scale was hypothesized to discriminate between DSM-IV
diagnoses classiﬁed as psychotic or non-psychotic and the substance-use scale is
hypothesized to discriminate between patients who are diagnosed with Substance-Abuse
Disorders and those that are not diagnosed with Substance-Abuse Disorders. In
addition, the Moods and Emotions Scale was hypothesized to discriminate between
those patients diagnosed with internalizing disorders (i.e. , Major Depression and
Anxiety Disorders) and those who are not diagnosed with internalizing disorders. The

Role performance scales (School, Home and Community) and Being With Others scale

3O

waehypodtesizedmdiscﬁmmatebetweenﬂrosechﬂdrarandadolescenmwhoare
diagnosed with externalizing disorders (i.e., conduct disorder, positional deﬁant
disorder, intermittent explodve disorder) and those who are not.

The FISCA also should be able to discriminate between patients with no
comorbidity, low comorbidity, and high comorbidity. Patients with greater
comorbidity should have a higher level of functional impairment.

Bredictixejlalidity: The last type of validity evaluated in this study is
predictive validity, using length of stay (LOS) as the criterion variable. Although LOS
is correlated with variables independent of the child's psychopathology, such as
insurance coverage (Patrick et al. , 1993) and age (Browning, 1986), it nevertheless has
been shown to be highly correlated with initial level of impairment as well (Gordon,
Jardiolin, & Gordon, 1985; Gordon, Vijay, Sloate, Burket, & Gordon, 1985). It is
hypothesized that a child who is more impaired should recover more slowly and thus

higher impairment in functioning will show a positive relationship with LOS.

HYPOTHESES

To summarize, a new multidimensional measure of child and adolescent
functional impairment, the Functional Impairment Scale for Children and Adolescents,
hasbeenpresented. Thepurposeofthis study wastoassessthedegreeofoverlap
between functional impairment as measured by the FISCA and patient symptomatology
and diagnosis and, as such, to evaluate the psychometric properties of this measure by
establishing concurrent, discriminant, and predictive validity. Hypotheses were as
follows:

1. The FISCA scales will demonstrate adequate concurrent validity in that:

Al. The Total FISCA score and each FISCA scale will correlate
signiﬁcantly with two other global measures of functional impairment,
the Children's Global Assessment Scale (Shaffer et al., 1983) and the
Columbia Impairment Scale (Bird et al. , 1993).

B1. The total FISCA score will correlate moderately and signiﬁcantly with a
total problem score on the Youth Self Report and Child Behavior
Checklist.

B2. Each of the individual FISCA scales will correlate signiﬁcantly with
measures of symptomatology, especially those assessing conceptually

similar problem areas. The literature indicates that somewhat lower

31

32

correlations are expected between the FISCA scales and corresponding
child—report measures of symptoms than with corresponding parent-
report measures of symptoms because of parent-child reporting

discrepancies.

2. The FISCA scales will demonstrate adequate discriminant validity as

operationalized by the following hypotheses:

A.

The FISCA will be able to discriminate between outpatient and inpatients
in that the mean FISCA score for inpatients will be signiﬁcantly higher
than the mean score for outpatients.

The FISCA will be able to discriminate between patients with no
comorbidity, low comorbidity, and high comorbidity. Patients meeting
criteria for a greater number of diagnoses will have higher functional
impairment scores than those meeting criteria for fewer diagnoses.

Each of the individual FISCA scales will be examined in relation to their
ability to discriminate between diagnostic groups. The most
conceptually similar FISCA scale was hypothesized to be the best

discriminator of a particular diagnostic group.

3. The FISCA scales will demonstrate adequate predictive validity in that:

A.

The FISCA will be signiﬁcantly and positively related to differences in

patient’s length of stay (LOS).

METHOD

SAMELE

The participants included both children and adolescents and their parents in an
inpatient setting and parents of children and adolescent outpatients. Five subsamples
were used to test various hypotheses. Sample I was used to test the relationship
between the FISCA and two other measures of functional impairment (the CIS and the
CGAS), and between the FISCA and diagnoses. Sample 11 was used to test the
relationship between the FISCA and parent report of symptomatology. Sample III was
used to test the relationship between the FISCA and the Youth Self Report (Achenbach,
1991b). Sample IV was used to test the relationship between the FISCA and adolescent
report of their alcohol and drug use. Sample V was used to test the ability of the
FISCA to discriminate between inpatients and outpatients.
Inpatients

Participants in the inpatient sample were 400 children and adolescents, ages 6 to
17, and their parents. The study consisted mainly of mothers, because mothers were
more likely to accompany their child to the hospital and to complete the measures. The
patients were consecutive admissions to two private Midwestem psychiatric hospitals,
one in Nebraska, and one in Michigan. Demographic information on the subsamples is

listed in Table 1. More than half of the total sample are Caucasian (53 %), however

33

3 4
this is not surprising because patients in both hospitals came primarily from rural areas
with largely Caucasian populations.

The Nebraska facility has a child and an adolescent unit and the Michigan
facilityhasthreeunits: onethattreatschildrenthatareunder ll yearsofageand
severely developmentally delayed older children; one that treats adolescents with more
”internalizing” problems; and one that treats adolescents with more ”externalizing"
problems. The ratio of males to females was expected to be approximately equal for
adolescents, but it was expected that for younger children, there would be substantially
more boys than girls. Children and adolescents hospitalized during the study period
who do not have a current FISCA were excluded. The diagnostic interview, parent and
self-report of symptoms data were collected from the Michigan facility. These data
were unavailable for the Nebraska facility. Analyses that utilize adolescent self-report
data will be subject to the following inclusion criteria: 1) the child must be at least 11
years old; 2)must have an IQ of 70 or greater or have no record of being
developmentally delayed; and 3) have no evidence of thought disorder or organicity.
GIMP-018

Participants will be 100 parents of children and adolescents, ages 6 to 17, who
are receiving outpatient services in one of two private Midwestern psychological
clinics. When subjects ﬁrst started receiving services at the outpatient clinic they were
asked to participate in the study. Both clinics also treat patients who were recently
discharged from the hospital. These patients were excluded from the study.

Demographic information on this sample is listed in Table 1. Outpatients who initially

35

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

9.8.1... 23:: 988...... 53.... 9.3.1... 38:... 98...... 33...! 6.8.1... «.8?! 9...... I... 2.8:... 8.8.38
8...“... .8...— 5...... .8...- 08.9. .8...— 38...... .85. 38.3. .8...— 833. H2...... 8......
9.2.. 1... 33.8. 6.288.... 3.?! 6.8.7.... 2.2:: 8.3.1... 3.2.... 6.8.1... 2..an 88.18.3372 8.88.8
8328...... 83m. 8.... 8.2.... "8.5. 2...... .8...— .8.%. ”8.3. 3...... .83. 2.32
8.33 333 8.38. 8.3 8393 833
33.385833. 33.383833. 3.383.833. 33.85.8330 8.. 33383823. 3.8.8583... 2.
logo—.3 «a... $0.033“. 8.8. I wrote—.3 o6... duo—.3 Reine—.3 «Ea. duo—.3 «Be.
88.3.8832... *8 333.8338 88 33.88338 38. a." .3338 a... .33....832... 88.33.3832...

8... 88 .3338 .8... :8 3.8.8 8... a... .3338 32.8 8... £8 .8... a." 3.8.8 .8... 3.13.338 8.32....
32.883.2388 32.88....85. 32.883.22.82 .38.»...85. 32.883.22.82 338.833.: 8......
839—. 8.33 8.30.. 02.8. 0333. 8.38.

3.883.833. 3.3-8 5.83.8 31.832833. 88.85833. 3.385.823. 3.385183. 8...
$06.53 Bee- &0 duo—.3 go. In duo—.3 8.8- lh.‘o=3 8.8. Induoaeo «Ee- .ouo=eo 8.8.
a... .32....8 323. E... .33....8 32.8 32 .33....8 .8..8 can .3338 32.8 up... .3338 32.8 *8 .33....8 32.8
.8... a... 3.2.8 .8... a: .3338 .8... a... .3338 8... a: .33....8 .8... a... 3.3.8 .8... no" .3338 828......
33882.8»... 3.883.833. 33883288.. 3383.38: 3388.....85. 3.8.82.8»... 2.3:
8...? A a. 8...? A an 8...... A a. 38.2. A a. 83... A an 8...? A .8
88.3.8. :8 88.8388 88 83.1.8.8 as... 88.8.8.8 8.... 89.83.8638 28.31.86." :8
89.8.38... no... 88.8.83. 2.. 85.838... 88 «8.8.892 88 «8.838.838 88.8.38... .38
can. v a: 38.". v a... .8... v E." 8...”. v *8 .8... v 32 8...". v a: 2.38.
88.35% i883... 3.5.2.8.... a. 53885.. no 5886...; .138...
8.35.3.8... 3.38.3.8... .33.... 8.3.8... 3.33338... 8.3382038... 33:838....
38 £3.35 3...: :8 3.832.... .88... a: .8812... 23.... an. .3383. 2a... u: .3? 3...... an 2.8.83. 33....
3.81.2338 8.3.1.863... 3.3.3.3689 33.1.3628 :qu 38.1.2582. 83......
3.85.8388 3.8.... .82.... 3.88.82. 8 3.88.838. .83....3... 3.83.823 .8...
.31... was... 2.9.... .87.. a... a... 3...: c... I... .8. I: 6?... 8.2-: Sun... 5?:
2.288: :28..- ....28...- 2.28.... :28...- .......8...u 8..
93383.3... >395... 838...... 5.358 :38... 3......”
3.8833153

36
came to the hospital requesting inpatient eare and were referred to the outpatient

clinics were excluded.

 

&_Eaul,_1995). The FISCA is a parent-completed questionnaire assessing functional

impairment in children ages 6 to 18 (See Appendix A). The FISCA measures a child
or adolescent's impairment in eight different areas: School, Home, Community,
Thinking, Being With Others, Moods & Emotions, Self-Harm and Alcohol and Drug
Use. Development of the FISCA is based in large part on criteria for level of
impairment identiﬁed by the Child and Adolescent Functional Assessment Scale
(Hodges, 1990). The FISCA is a questionnaire that can be ﬁlled out by a wide variety
of informants including parents in order to provide information needed to assess the
presence of impairment criteria like those deﬁned by the CAFAS. The FISCA uses a
multiple choice format with speciﬁc responses. These responses are keyed to a
particular criteria so that scoring is automatic. This is in sharp contrast to scoring of
previous measures, including scoring of the interview accompanying the CAFAS,
which mostly require clinical judgement.

For each area of impairment the CAFAS scale had criteria for mild, moderate
and severe levels of impairment. The FISCA also has criteria for mild, moderate and
severe levels impairment in each area of impairment (See Appendix A). The
development of the FISCA criteria entailed reorganizing, modifying and in some cases

deleting CAFAS criteria, as well as adding a few new criteria. Frank & Paul (1995)

37

deleted CAFAS criteria that used prior history or mental illness to rate the child as
more impaired. The authors also modiﬁed criteria that were overly detailed, redundant
or ambiguous and added relevant criteria dealing with eating problems. The revised
criteria were organized to create the FISCA scale (See Appendix A). For each criteria,
the authors developed objective questions to assess whether the child would meet each
of the various criteria. For instance, for the criteria ”set ﬁres with malicious intent”
theparentisaskediftheirchildsetﬁresonpurposetodesmypropertyorhurtpeople.
These questions were combined to create the FISCA questionnaire. The FISCA is
scored using a scoring key that matches items or combinations of items to relevant
criteria.

W. The
CGAS is a measure of severity of functional impairment which was adapted from the
Global Assessment Seale developed for adults by Endicot et al., (1976). The measure
has a range of 1 through 100 and provides anchor point descriptions of behavioral and
emotional functioning for each decile. Scores above 70 indicate functioning in the
normal range. It was designed for use with children 4 through 16 years of age. The
scale was designed to be scored by a trained clinical interviewer. Bird et al. (1987)
reported satisfactory test-retest stability over a 19 day period (.85). The CGAS was
found to discriminate between outpatients (65.4) and inpatients (46.0) (Shaffer et al. ,
1983). The correlation between the CGAS and the Conners Abbreviated Scale was
-.25 (p > .05, df= 17) in an outpatient sample (Shaffer et al., 1983). Green et al.
(1994) found that the CGAS correlated with Activity Competence (.26), Full Scale IQ

(.43), Social Relatedness (.33 for psychiatrist ratings, and .39 for milieu staff ratings).

38

The CGAS has also been found to correlate (r=-.80, p < .01) with the total score from
a structured diagnostic interview, the Child Assessment Schedule (Hodges, Kline,

Stern, Cytryn, & McKnew, 1982).

 

C18 is a 13-item scale that can be administered by a lay interviewer to provide a global
measure of impairment. The scale was developed to tap four major areas of
functioning; interpersonal relations (e. g., How much of a problem do you think he/she
has with getting along with other kids his or her age?) certain broad areas of
psychopathology (How much of a problem do you think he/she has with her behavior at
home?); functioning at school or work (How much of a problem do you think he/she
has with his/her schoolwork?) and use of leisure time (How much of a problem do you
think she has getting involved in activities like sports or hobbies?). Factor analyses
reveal that the scale is measuring one dimension. As a result, this study will use only
the total score. Parents use a Likert scale ranging from 0, ”no problem", to 4, ”a very
big problem” to respond to each item; the total score can range from a minimum of 0
to a maximum of 52. Bird et al. (1993) report good test-retest reliability for parent-
report (Chronbach's alpha=0.89). The parent CIS correlates signiﬁcantly with a
therapist's rating (r=-0.56), with grades in school (.45).

MmresniSyptomatology

W The CBCL
is a well-known, standardized measure of parent perceptions of behavior problems of

children ages 4 to 18. It contains 118 speciﬁc behavior problem items and 20 social

competence items. Parents are asked to rate behaviors that have occurred during the

3 9
last six months using a 3-point scale of 0 (not true), 1 (somewhat true), or 2 (often
true). The checklist comprises 8 factor-based I'narrow-band" syndromes and two
global ”broad-band” syndromes, which were developed using second-order factor
analysis of the narrow-band syndromes. The narrow-band syndromes are labeled
Withdrawn, Somatic Complaints, Anxious/Depressed, Social Problems, Thought
Problems, Attention Problems, Delinquent Behavior, and Aggressive Behavior. The
broad-band syndromes are labeled Internalizing, which includes the narrow-band
Withdrawn, Somatic Complaints, and Anxious/Depressed scales, and Externalizing,
which includes the Delinquent Behavior and Aggressive Behavior syndromes. In
addition, all the items (but two) can be added to compute a total behavior problem
score.

Social competence items require the parent to list the sports, hobbies, activities,
clubs and jobs that the child is involved in and to rate their skill in each compared to
other kids their child's age. In addition, the parent reports on how many friends the
child has, how much time they spend with friends, the child's ability to get along with
others, and his or her academic achievement. The social competence items form three
scales, Activities, Social, and School. The scores from each of these three scales can
be summed into a Total Competence score.

The CBCL manual provides evidence for the reliability of the CBCL. The
reliability data was analyzed separately for children aged 4 to 11 and 12 to 18, and for
boys and girls. The internal reliability of the narrow-band scales was assessed using
Cronbach's alpha and ranges from .62 (Thought Problems, for boys, ages 4-11), to .92

(Aggressive Behavior for all groups). Cronbach' s alphas are at least .89 for the

4O

Internalizing Scale and at least .93 for the Extemalizing scale for the total sample.
Alphas for the Social Competence scales were lower than for the other scales, ranging
from .42 (Activities for boys ages 4-11) to .64 (Total Competence for boys 12-18).
Test-retest reliability over a one week period was .89 for the Behavior Problems scales
and .87 for the Social Competence scales.

The CBCL manual also provides evidence for the validity of the CBCL. The
CBCL scales have shown modmte to high correlations with similar scales generated
from the Conners Parent Questionnaire (Conners, 1973), the Revised Behavior
Checklist (Quay & Peterson, 1983) and the Werry-Weiss-Peters Activity Scale (Mash
& Johnston, 1983). The CBCL has been shown to discriminate between referred and
nonreferred children (Achenbach, 1991b).

WW). The YSR is a widely used self-
report measure designed for 11 to 18 year olds. The YSR contains 103 speciﬁc
behavior problem items and 17 social competence items. The YSR was developed to
obtain adolescents' views of their own behavior problems in a manner that would
facilitate comparison with parental report on the CBCL. The YSR and the CBCL have
89 items in common. The YSR is scored the same way as the CBCL and contains the
same narrow-band and broad-band syndromes. However, the Social Competence items
only form two scales, Activities and Social.

The YSR manual provides evidence of the reliability of the YSR. The internal
reliability of the narrow-band scales was assessed using Cronbach's alpha and ranges
from .59 (Withdrawn for both sexes) to .90 (Anxious/Depressed for girls). Cronbach's

alphas for the Internalizing Scale are .89 or greater and for the Extemalizing scale are

41

.89 for both sexes. Alphas for the Social Competence scales were lower than for the
other scales, ranging from .32 (Activities for girls) to .60 (Social for girls). Test-retest
reliability over a one week period was .72 for the Behavior Problems scales and .76 for
the Social Competence scales.

The YSR scales have shown moderate correlations with similar constructs
derived from the CBCL parent and teacher forms (Stranger & Lewis, 1993). In
addition, the YSR discriminates between children referred for mental health services

and nonreferred children (Achenbach, 1991b).

 

AAIS is a 13 item self-report measure of the adolescent's alcohol use and involvement.
The items assess the quantity and frequency of alcohol use, negative consequences
resulting from excessive drinking (e. g. , hangover, drunken driving), and contextual
factors indicative of the amount of alcohol involvement (e.g. , time of drinking, reasons
for drinking, age of ﬁrst drink, etc.). The items are totaled, with a higher score
indicating more serious alcohol involvement. Mayer and Filstead (1979) reported two
week test-retest reliability at .89 for controls and .91 for patients at an alcohol
rehabilitation center and patients at an inpatient psychiatric hospital who misuse
alcohol. They found no signiﬁcant difference between boys’ and girls’ scores on the
measure. They did ﬁnd that in both groups the mean score at each year of age (13-18)
was signiﬁcantly different (p < .01) except between the ages of 13 and 14. The mean

total score for the control group was 19 and the mean total score for the experimental

group was 58.

 

unpublishedjnsmrment). This is a 41 item self-report measure of adolescent drug use
and involvement. It uses a 9-point Likert scale to assess how often the respondent had

used marijuana, hash and 10 "hard” drugs (LSD, psychedelics, cocaine,
amphetamines, quaaludes, barbiturates, tranquilizers, heroin, and other narcotics). The
adolescents also report on negative consequences associated with drug use (e.g., “lost
friends“; ”been ﬁred"; ”had to go to hospital“). The responses were used to compute a
total drug use and total drug consequences. Reliability and validity of this scale are
currently being studied (Fitzgerald, personal communication). This measure has been

used successfully in studies of adolescent inpatients (Frank & Burke, 1992; Frank &

Poorman, 1993).

 

diagnostic interview, the K-SADS-P, will be used to assign diagnoses to a subsample of
75 patients (25 from each of 3 units: a unit that treats children under age 11 and
severely developmentally delayed older children; one that treats adolescents with more
"internalizing” problems; and one that treats adolescents with more ”externalizing”
problems).

The K-SADS-P assesses the patient's functioning and symptoms for a variety of
psychiatric disorders, including depressive disorders, mania, psychosis, anxiety
disorders, attentional disorders, conduct disorders and substance abuse disorders.

Test-retest reliability coefﬁcients for the K-SADS are generally moderate to

4 3
high, with anxiety disorders showing the poorest reliability (Chambers et al., 1985).

The K-SADS is designed for use with children ages 6-18. While the validity of
the K-SADS at different ages has not been evaluated empirically, evidence has
suggested that stability of child self report information increases with age (Edelbrock,
Costello, Kalas, Dulcan, & Conover, 1985).

Hodges, McKnew, Burbach & Roebuck (1987), reported that there was good
concordance (ranging form .59 to .65) between the CAS (Hodges, 1982) and the K-
SADS (Puig-Antich & Chambers, 1978), although low concordance was found for
anxiety disorders. This study and other ﬁndings in the literature suggest that the poor
concordance for anxiety disorder may reﬂect broader disagreement as to the criteria
used to diagnose anxiety disorders (Hodges, 1987; Chambers et al., 1985; Costello et
al. , 1984).

Green et al. (1987) examined the K-SADS in relation to the Diagnostic
Interview Schedule for Children (DISC; Costello, 1984). Agreement on diagnom
based on the K—SADS and the DISC were moderate and signiﬁcant, with the exception
of overanxious disorder which yielded nonsigniﬁcant agreement.

ImgthnLStay

Length of stay will be obtained from howital discharge records that list the
patient's admission date and discharge date.
Demographiclnformation

Demographic information (i.e. , age, sex, mother education and occupation,
father education and occupation, income and ethnicity) were obtained from a short

demographic questionnaire. This questionnaire was developed by Frank (1995) to

44

assess basic demographic information for all patients seen at both the inpatient and
outpatient facilities. Patients are asked to state how old their child is, and to report
what their child's gender and ethnicity and the family income is. Income is rated on a
scalefromoaessthan8,000)to6(morethan 100,000). Parentsareaskedtoindicate
the highest educational degree they have received. This is rated on a scale from 0
(some high school) to 7 (graduate or professional degree). The occupation scores were
rated using the Duncan coding system (Mueller & Parcel, 1981). Higher numbers
indicate jobs associated with a higher socio-economic status.
Procedures

This study involves two separate procedures; one for an inpatient group, and
one for an outpatient group.
Inpatients

Parents of inpatients in the Nebraska facility also only completed the FISCA
and Demographic Questionnaires. Michigan participants completed the FISCA and
demographic questionnaire. In addition, subsamples of the Michigan group completed
other measures. One subsample of Michigan parents also completed the CBCL
(Achenbach, 1991a). Another subsample of Michigan parents also completed the
diagnostic and impairment measures. A third sample of Michigan patients completed
the Youth Self Report (Achenbach, 1991b). A fourth subsample of Michigan patients

completed the Drug and Alcohol Questionnaires.

Mamrmﬁompletedatlntake

Dmmgmpthmﬁonnaimﬂanij). This measure was administered to
the parents when they ﬁrst brought their child to the hospital. This was administered in

45

 

&_Paul,_1995). 'I‘hefirstmeasurethatparticipantswereaskedtocompleteisthe

FISCA. Both the Nebraska and Michigan participants completed the FISCA. The
FISCA is used as a part of the intake procedure. As with the outpatient sample, the
FISCA is first introduced with a time-line. The FISCA was designed to help staff
decide on the appropriateness of a hospital admission, however the FISCA was
introducedduringthestudyperiodandasaresultitwasoftennotscoredoreven
consulted until after admission. In addition, during the study period, staff often scored
the FISCA incorrectly. As a result, the author used a computer program to rescore all
FISCA’s used in this study.
WWW Once a
decision to admit had been made, staff provided parents with a parent assessment
packettoﬁlloutinthewaiting roomorathome. TheCBCLwasapartofthis
assessment packet. Parents returned the packet either at admission or upon their next

visit to the hospital.

 

being conducted as a part of this study, as well as, another study at the Michigan

facility. At admission, the purpose of the study was explained to the parents. If the
parents agreed to participate, they were asked to sign an informed consent (see

Appendix C) and the child was asked to sign an informed assent (see Appendix D).

4 6

Participating parents received a $10 gift of appreciation after completing the interview.
Since parents complete a lengthy protocol at admission and scheduling timely post-
admission contact was difﬁcult, the K-SADS-P interviews were conducted over the
telephone within 48 hours of admission. Although no studies have assessed the
reliability of this procedure, Hammen (1988) conducted telephone interviews for
follow-up diagnoses, and found that subjects responded in a very similar manner over
the phone as they did in face-to-face interviews. Within 48 hours after completing the
interview with the parent, the interview with the child and adolescent was conducted in
person. Interviews were conducted by the author and one other graduate student in
clinical psychology who underwent rigorous training in the interviewing and scoring
procedure by an experienced interviewer. To determine summary diagnoses based on
both parent and child report, the two interviewers independently reviewed the scoring
protocol and arrived at the ﬁnal diagnoses. When there were differences they reviewed
the diagnosis together and arrived at a consensus. Twenty parent and twenty child
interviews were audiotaped and scored independently by both interviewers to provide a
coefﬁcient of interrater reliability.

IheColnmhialmpainnenLScale:ParentEonn.(ClS;.Bird_etal.,_1993). The
CIS was administered over the phone to the parent at the end of the K-SADS interview.
The parent was asked each question and presented with the ﬁve possible options. The
interview was done within 48 hours of the child's admission.

Ihe_Chﬂdrenls_Glohal_AssessmenLScale(CGAS;_Shaffer_eLal.,_1983). In this
study, the CGAS was completed by the author and one other graduate student in

clinical psychology who underwent training in the rating procedure. The students read

47

vignettes and rated the CGAS scores. When the raters disagreed, they reviewed the
vignette together and arrived at a CGAS score. The CGAS was completed using
information obtained during the K-SADS interview with both the parent and the child.
Twenty parent and twenty child interviews were audiotaped and CGAS scores were
independenﬂyratedbybothratersonacommonsetofprotocolstoprovidea
coefﬁcient of interrater reliability.
Chﬂdzmoereasures

The Youth Self Report (YSR; Achenbach, 1991b), Adolescent Alcohol
Involvement Scale (AAIS; Mayer & Filstead, 1979) and the Michigan Drug Use
Questionnaire (Zucker, Noll, & Fitzgerald, unpublished instrument) were completed by
patients who are 11 to 17 years old. The patients had an IQ of 70 or greater or had no
record of being developmentally delayed and they had no evidence of thought disorder
or organicity. These measures were administered within 5 days of the child or
adolescent's admittance to the hospital. Children who stayed less than ﬁve days may
not have been able to complete these measures. These measures were given as part of a
group testing situation in which trained undergraduate psychology students were
available to assist in reading and to answer questions.
mantis-Ms

All outpatients in this sample came to the clinic to receive outpatient care and
none requested inpatient care. Outpatients who initially came to the hospital requesting
inpatient care and were referred to the outpatient clinics were excluded, so that the
ability of the FISCA to discriminate between outpatients and inpatients could be more

clearly examined. Parents of children in the outpatient sample only completed the

48

FISCA and the demographic information sheet. The other data was not available for
this sample. The FISCA was used as part of the initial assessment. The FISCA was
ﬁrst introduced with a time line. The goal was to allow the parent to clearly orient him
or herself in time so that the child's functioning can be accurately recorded or described
for a speciﬁed time period. In particular, parents were asked to concentrate on speciﬁc
deﬁcits in functioning during the past three months (See Appendix B). Parents were

asked to complete the FISCA in the waiting room before their ﬁrst appointment.

RESULTS

EmliminaQI—Analm

Before testing each of the hypotheses, I ran preliminary analyses to examine
whether any of the demographic variables could confound the results. These analyses
were run for each of the 5 samples used to test the various hypotheses. For the most
part, the demographic variables assessed in this study (i.e. , father occupation, father
education, mother occupation, mother education, income, age, gender and ethnicity)
were related to both the independent and dependent variables assessed for any given
hypotheses. The one exception was in Sample V (the Inpatient-Outpatient sample). In
this sample age was related to total FISCA score as well as inpatient vs. outpatient
status (i.e., level of care). As a result, age was controlled for in all analyses using the
inpatient-outpatient sample (i.e. , tests of Hypothesis 2A).
W

Hypothesis 1A states that the concurrent validity of the FISCA can be
demonstrated by its signiﬁcant relationship with two other measures of functional
impairment: the Columbia Impairment Scale (CIS, Bird et a1. , 1993) and the Children’s
Global Assessment Scale (CGAS, Bird et al. , 1987). Sample I (the diagnostic sample),
consisting of data from 75 Michigan inpatients was used to test this hypothesis. This

sample consisted of 53 mothers, 4 step-mothers, and 10 “others” (i.e., aunt,

49

50

grandmother and foster mother). Interrater reliability for the CGAS, established by 2
independent raters on 22 protocols, was very satisfactory (r=.85, p< .001).
Correlations between the FISCA and the CIS, and between the FISCA and the
CGAS, supported the hypotheses (Sm Table 2). Higher scores indicative of greater
impairment on the CIS correlated with higher scores (i.e., more impairment) on the
FISCA. The FISCA and the CGAS also correlated signiﬁcantly, with lower CGAS
scores (indicative of more impairment) associated with higher FISCA scores. Among
the FISCA subscales, School, Self-Harm, and Alcohol and Drug Use correlated
signiﬁcantly with the total CIS score, whereas, the School, Thinking, Mood, and
Self-Harm FISCA subscales correlated signiﬁcantly with the CGAS. In short, the
FISCA , as expected, was related to two other measures of functional impairment, the

CIS and the CGAS. These ﬁnding support the concurrent validity of the FISCA.

HypotheseLlBandJC
CBCLSample

The total FISCA score was also expected to correlate in meaningful ways with
scores on the Child Behavior Checklist (Hypothesis 1B). Sample 11, used to test this
hypothesis, consisted of 150 mothers, 21 fathers, 4 step-mothers, and 20 “others” (i.e.,
grandmother, aunt, foster-mother and legal guardian) of 195 children and adolescents at
the Michigan inpatient facility. The FISCA was hypothesized to correlate with the total
CBCL score and the internalizing and externalizing CBCL scales. Correlational
analysis supported this hypothesis. The analyses indicated that the total FISCA score

correlated signiﬁcantly with both the CBCL internalizing (r=.25, p< .000) and CBCL

51

Table 2

C18 CGAS
SCHOOL .40" -.24‘
HOME -. 10 .15
COMMUNITY .17 -.16
THINKING .03 -.34"
BEING WITH OTHERS . 18 -.10
MOOD .15 -.24‘
SELF-HARM .29‘ -.34‘
ALCOHOL & DRUG .23‘ -.04
TOTAL FISCA .41" -.41"

' ps .05
" pg .001

52

externalizing (r=.51, p< .000) scores as well as, with the total CBCL score (r=.21,
p< .000); notably the relationship between the total FISCA score and the externalizing
scale was somewhat stronger than the relationships between the total FISCA score and
either the CBCL internalizing scale or the total CBCL problem score.

A third hypothesis (1C) was that each of the individual FISCA scales would
correlate with CBCL narrow-band scales measuring conceptually similar problem areas.
Overall, most of the expected correlations between the CBCL and the FISCA were
signiﬁcant. These ﬁndings support the concurrent validity of the FISCA.

A total of 19 correlations were expected to be signiﬁcant (these associations are
shown in bold numbers in Table 3). 18 of the predicted relationships were signiﬁcant
at p< .05. By chance alone only one of 19 correlations should be signiﬁcant at the
.05 level; however, as many as 16 of the predicted relationships were still signiﬁcant
using a more conservative alpha of .003 established according to the Bonferonni
correction procedure (Dunn, 1961). The one exception was failure to ﬁnd a signiﬁcant
relationship between the FISCA school scale and the CBCL aggression scale. The
additional exceptions were the relationships between the FISCA mood scale and the
CBCL somatization scale and between the FISCA school scale and the CBCL
externalizing scale.

Several unpredicted relationships between the FISCA subscales and the CBCL
scales also were statistically signiﬁcant (See Table 3). These correlations typically
made theoretical sense but were somewhat lower than those predicted on an a priori

basis.

53

 

CICLTotal

'95 .05
‘ps .001

School

.13

.14

.19'

.11

.18‘

.01

a,

‘Hypodsesiaedconeletionsareinbold

MALES

C 'I W
-.10 .15
-.06 .07
-.01 .21‘
.02 .31'
.11 .39’
.12 .30‘
.35‘ .16‘
14' .35‘
-.05 .19'
31‘ .32‘
.11 .11

.23‘

.28"

.18

.30’

.31“

.53.

.15

.51“

.19'

.13

.17‘

.12

.21'

.21"

.13

.11

.26"

.19‘

.30‘

.30‘

.47‘

.42‘

.51‘

.28‘

54

YSlLSample

The FISCA was also expected to correlate in meaningful ways with scores on
the Youth Self Report. These ﬁndings partially support the concurrent validity of the
FISCA. Sample 111, consisting of 192 children from the Michigan inpatient facility,
was used to assess the relationship between adolescent patient’s reports of symptom
behaviors and the FISCA. Only the relationship between the FISCA total score and the
YSR externalizing scale was signiﬁcant (r=.21, p< .01).

The total FISCA did not correlate signiﬁcantly with either the total Youth Self
Report (YSR) score (r=.06) or the YSR internalizing scale (r=-.05). These ﬁndings
indicates that the FISCA is more related to child report of externalizing problems than
to internalizing problems.

Each of the FISCA subscales were hypothesized to relate to conceptually similar
problem areas on the YSR. Overall, child reports on the YSR scales were less likely
to be associated with the FISCA scales than parent reports of child problems on the
CBCL, even though the number of signiﬁcant FISCA-YSR correlations were greater
than what would be expected by chance. A total of 19 correlations were expected to be
signiﬁcant (these associations are shown in bold numbers in Table 4). By chance only
one would be expected to be signiﬁcant. In fact, 6 of the 19 predicted correlations
were signiﬁcant. Five of the six signiﬁcant relationships (p < .05) are with the School,
Home and Community scales. A few meaningful non-hypothesized relationships also
were signiﬁcant. Overall, these ﬁndings provided only weak support for the

concurrent validity of the FISCA.

55

Table 4
(zomIan'ms mam EIS‘ : A smlm and XSB smlest
WALES
m School Home Community Thinking Being will Moods & Total FISCA
mam others Emotions
Withdrawn -.04 -.l4‘ -.24‘ -.06 -.07 .03 -.09
M .01 -.09 -.l9' .04 .04 «N -.08
Coqlehls
Anxious] -.11 -.21 -.25" .05 -.03 .N -.05
Depressed
Social 02 - 04 -.06 09 .05 - 05 - 01
Problem
Thought - 06 -.05 - 04 15' - 05 06 09
Problems
Arsenic- .07 -.04 -.09 .13 .08 .05 .12
Problem ‘
Dem .19‘ .15' .19' .03 .03 -.02 .25
Behavior
Agree-vs .16' .U .02 12 .12 01 16
Behavior
hes-nelizir' -.07 -.18 -.26 .03 .03 .05 -.05
Extemalizing .19' .11 .09 .09 .09 -.(X) .21‘
YSR Total .03 -.08 -. 14 .08 .08 .05 .06
" ps .05
'ps .MI

I'Hypotllesizerlcorrelationsar'eirrbold

56

We
The FISCA was also hypothesized to correlate with adolescent patients’ reports

of their own drug and alcohol use. Sample IV, consisting of 200 adolescents from
Michigan was used to test the hypothesis that the Alcohol and Drug use FISCA

subscale in particular would be related to youth self-report of alcohol and drug use.
Results substantial this hypothesis; and as such strongly supported the FISCA’s
concurrent validity. The Alcohol and Drug use FISCA subscale did correlate
signiﬁcantly with all 5 alcohol and drug use variables from the youth self report
measures of substance use (see Table 5). The total FISCA score also correlated with
youth self-report of substance use. Notably, of 35 possible correlations between the
other FISCA subscales and the adolescents scores for the alcohol and drug variables,
only 5 were signiﬁcant at p< .05 (2 would be expected by chance).
HypothesisZA

The FISCA was expected to discriminate between inpatients and outpatients.

Sample V, consisting of 50 outpatients and 50 inpatients for the Michigan facility and
50 outpatients and 50 inpatients for the Nebraska facility were used to test whether this
prediction (Hypotheses 2A). Reporters in this sample consisted of 148 mothers, 25
fathers, 3 step-fathers and 24 ”others” (e.g. , grandmother, foster mother, foster father
and legal guardian). Because age was associated with both level of care ( M for
outpatientswas 12.05, M for inpatients was 13.52, t(l96)= 3.39, p< .05) and the total
FISCA score (older patients were more impaired, r=.25, p< .01) analyses testing this
hypothesis included age as a covariate. As expected inpatients scored signiﬁcantly

higher on the FISCA (M= 138) than outpatients (M= 103.2) even after controlling for

57

ha.

.2.

bu.

.nN.

<09“—

3’-

.3.

39—. i3

.3.

2.-

Z.

S.

Eel-20m

no.

2.

2.

Q.

8.-

S.-

8.-

2.-

.3.-

3.

.2.

.2.

I.

on.

32. a a. 5388 83858»: .

2.

no.

no.

.3.

.2.

Q.

2.

“8.8.

8.8.

a:
mac 2.:

.5
3.. €29

8: .8
25.3:

SD
.9302

«co—Eire—
834.
.332

 

58

age (E(1,195)=33.7, p< .01). Additional analyses indicated that differences in the
FISCA score associated with level of care did not differ as a function of facility
location (Michigan vs. Nebraska) (E(1,194)=.15, p=n.s.).

Two discriminant analyses examined which FISCA scale or combination of
FISCA scales would best discriminate between outpatient and inpatients. Results from
both analyses supported the research prediction. Findings for the ﬁrst analysis which
ﬁrst controlled for age and then tested all of the FISCA scales in a stepwise fashion are
shown in Table 6. After entering age, the Home, Self-Harm, and Alcohol and Drug
subscales met criteria for entry into the analysis and correlated most highly with the
discriminant function. This function had a canonical correlation of .59 (p< .001) and
correctly classiﬁed 70.2 % of the participants into their actual level of care.

In a second analysis only age and the total FISCA score were entered into the
discriminant analysis, resulting in a discriminant function with a canonical correlation
of .44 (p< .001). As can be seen in Table 7, classiﬁcation results for the FISCA
subscales were very similar to results for total FISCA score. Both functions were able
to successfully categorize between 66 and 69 % of the sample into the appropriate level
of care; however the total score was able to predict inpatient status somewhat better
than the subscales whereas the subscales were somewhat better at predicting outpatient
status than the total score.

Hmhesiﬂﬁ
Hypothesis 2B, arguing that the total FISCA score would be able to discriminate

tested with Sample I consisting of 75 parents and their children from the Michigan

59

 

eﬁer fuel up amt-on Hm
Coefﬁcieu

AGE 36 54

SCHOOL . 16

HOME 53 66

COMMUNITY 1' .39

TWO a .14

name wrrH . .24

OTHERS

moons a =- .24

enormous

SELF-HARM .48 .50

ALCOHOL a DRUG .39 .56

’SceledidnotmeetF-vehecrkerieforemryimotheemlyses.

 

Age, Home, Sherm & Alcdrug Age, Total FISCA
D' . . F . D' . . F ion
I of Cases Predicted Predicted I of Cases Predicted Predicted
Oinpetierls 98 58 40 98 67 31
59.2% 40.8% 68.4% 31.6%
W I“) 27 73 100 30 70

27.0% 73.0% 30% 70%

60
facility. To test this hypothesis, patients were divided into several (overlapping)
diagnostic groups based on results from the diagnostic interview (K-SADS; Puig—
Antich, 1995). The Internalizing group consisted of individuals who had a diagnosis of
Panic disorder, Major Depressive Disorder, Depressive Disorder Not otherwise speciﬁed,
Separation Anxiety Disorder, Dysthymia, Generalized Anxiety Disorder or Phobic
Disorder (N =56). The Extemalizing group consisted of individuals with diagnoses of
Oppositional Deﬁant Disorder, Attention-Deﬁcit Hyperactivity Disorder, Intermittent
Explosive Disorder, or Conduct Disorder (N =66). The SubstanceAbuse group
consisted only of individuals with a Substance Abuse diagnoses (N = 17). The Thought
Disordra group consisted of individuals who had a diagnoses of Schizophreniform or
Schizoaffective Disorder (N =3). Individuals could be in more than one diagnostic
group. Interrater reliability for the diagnostic interview established for 22 protocols
was satisfactory. Two raters had perfect agreement as to the presence or absence of
Internalizing and Extemalizing diagnostic categories (kappa=1). The raters were also
able to reliably diagnose substance disorders (kappa=.90). Reliability for thought
disorders was not possible to examine since none of the 22 patients in the reliability
sample had a diagnoses of thought disorder.

Hypothesis 2B stated that patients with greater degrees of comorbidity would
have higher FISCA scores. The data supported this prediction. Preliminary analyses
indicated that mean FISCA total scores for individuals with 3,4, or 5 diagnoses were
very similar; hence these individuals could be conceptualized as a single ”High
comorbidity“ group. In this sample, 5 patients had no comorbidity, 18 had low

comorbidity (2 diagnoses) and 52 had high comorbidity (3 or more diagnoses).

61

A Sex x Comorbidity Group ANOVA of the total FISCA score resulted in a
non-signiﬁcant F for the two—way interaction between sex and comorbidity. Hence the
ﬁnal test of Hypothesis 2B used a 1-way analysis of variance.

Results of this analysis indicated that the total FISCA scores could in fact
discriminate between patients with different levels of comorbidity (E(1,72)= 8.2,
p< .05). Mean total FISCA scores for each comorbidity group were 110.0
(SD=33.91) for the single diagnoses group, 132.78 (SD=34.09) for the low
comorbidity group and 156.2 (SD=28.06) for the high comorbidity. The linear trend
analysis also resulted in a highly signiﬁcant E(l,72)= 10.8, p< .002. A post-hoe
comparison test using Newman Keul's test indicated that the single diagnostic group
differed signiﬁcantly from high comorbidity sample (p< .05).

Hypothmislc

Each of the individual FISCA scales were hypothesized to be able to
discriminate between diagnostic groups. Speciﬁcally the most conceptually similar
FISCA scale was hypothesized to be the best discriminator of a particular diagnostic
group. Sample I was used to test Hypothesis 2C. In general, the FISCA scales were
relatively successful at classifying cases into diagnostic groups. With the exception of
the Moods and Emotions scale and internalizing diagnoses, these ﬁndings support the
discriminant validity of the FISCA.

The hypothesis that the Moods and Emotions scale would be able to
discriminate between patients with and without an internalizing diagnoses was not

supported. A discriminant analysis using only the Moods and Emotions subscale failed

62
to discriminate between these patients. A stepwise discriminant analyses indicated that
the School subscale for the FISCA alone among the various subscales was able to best
discriminate between these groups; however the Moods and Emotions scale correlated
.37 with this function (See Table 8). The School scale resulted in more true positive
classiﬁcations than true negative classiﬁcations (See Table 9).

Since the Mood Scale alone was unable to discriminate between individuals with
and without an internalizing diagnoses, additional analyses of the scale’s ability to
discriminate between speciﬁc internalizing diagnoses was conducted. The mood scale
was not able to discriminate between individuals with and without a diagnoses of
Depression or those individuals with and without an Anxiety Disorder. However only
7 of the 58 individuals diagnosed with Internalizing disorder did not also have an
externalizing diagnosis.

The hypothesis that the Being With Others, School, Home and Community
subscales would be able to discriminate between patients with and without an
externalizing diagnoses was partially supported by the data. A discriminant analysis, in
which the School, Home, Community and Being with Others subscales were entered
simultaneously into the analyses, resulted in a function with a canonical correlation
coefﬁcient of .39, p < .05. School (canonical function coefﬁcient= .63, correlation
with function = .73) and Being with Others subscales (canonical function= .63,
correlation with function = .76) were better discriminators than either the Home
subscale (canonical function = .24, correlation with function = .23) or Community
subscale (canonical function = .06, correlation with function = .24). The School, Home,

Community and Being with Others function accurately predicted group membership for

 

Rhoda

m3

85.9

ﬁnd.»

*8.
E88.—

2

an

c3852

gag

«snafu

.Rmﬁm

ﬁe.»

R—.3

$5.49
28cc.—

3

~ _
c3852

3

$3.60

$4.6m

Re.“

2

e330
282.80
830

89.20
0%.“sz
83m

2.8.—
9.3

26¢sz
3:.

328m
03,—.

 

65

79.45% of Sample I.

In a second stepwise discriminant analyses using all of the FISCA scales, only
the Being with Others and the School subscales for the FISCA were able to
discriminate between individuals with and without externalizing diagnoses (See Table
8). Both Community and Home had very low correlations with this function, but
interestingly, the correlation between the FISCA Moods subscale and this function was
relatively substantial (.37) (See Table 8). The discriminant function for this analysis
successfully predicted 78.1% of group membership.

The hypothesis that the Alcohol and Drug scale would be able to discriminate
between those individuals with and without Substance Abuse diagnosis was supported.
A discriminant analysis in which only the Alcohol and Drug subscale was entered in the
analysis was able to signiﬁcantly discriminate between those patients with and without a
Substance Abuse diagnosis. A stepwise discriminant analysis, simultaneously entering
all FISCA scales still showed that only the Alcohol and Drug FISCA scale was able to
discriminate between the groups (See Table 8). This scale was able to accurately
classify 84.9% of the Diagnostic sample as in or out of the Substance Abuse diagnostic
group (See Table 9).

Finally, the hypothesis that the Thinking scale on the FISCA would be able to
discriminate between patients with and without Thought Disorder Diagnoses was
supported. A discriminant analysis using only the thinking scale indicated that the
drinking scale alone was able to correctly classify 64.4% of this sample. The Moods
and Emotions function had a canonical correlation of .24 with this function. A

stepwise discriminant analyses showed that the Thinking subscale and also the Being

66

with Others subscale together discriminated between the groups (See Table 8). The
discriminant function correctly predicted group membership for 84.9% of the sample.
The correlation with the total function was somewhat higher for the Being with Others
scale than the Thinking scale.

Hypothesis}

Hypothesis 3 stated that scores on the FISCA would predict length of stay for
inpatients. Sample VI consisting of 396 patients, 346 from the Michigan facility and
50 from the Nebraska facility was used to test this hypothesis. This sample consisted of
285 mothers, 7 step-mothers, 53 fathers, 7 step-fathers, and 44 “others” (e.g., aunt,
grandmother, and legal guardian). Notably, the location of the facility was related to
both length of stay (t =-2.23, p< .05) and total FISCA score (t=-2.6, p< .01). In
addition, regressing LOS on the FISCA X location interaction variable (after
controlling for the two main effects) resulted in a signiﬁcant interaction (R2 = .05,
p< .05). Length of stay was more strongly related to the total FISCA score for the

Nebraska facility (r=.47, p < .01) than for the Michigan facility (r=.13; p< .05).

DISCUSSION

The aim of this study was to validate the Functional Impairment Scale for
Children and Adolescents (FISCA) by establishing its concurrent, discriminant, and
predictive validity. As such, the study assessed the degree of overlap between
functional impairment as measured by the FISCA, and patient symptomatology and
diagnoses. This discussion will summarize and comment on the validity ﬁndings and
then discuss what the data suggests as to relation between functional impairment and
symptomatology, as well as functional impairment and diagnoses.
CnncurrmtXalidity.of.EISCA

Relation_to_other_measures_o£ﬂrnctional_irnpairment. One way of examining
the FISCA’s construct validity was by assessing its association with two other measures
of functional impairment: the Columbia Impairment Scale (CIS) and the Children's
Global Assessment Scale (CGAS). As expected, the FISCA total score was correlated
signiﬁcantly with both the CIS and CGAS. Patients showing more impairment on the
FISCA showed more impairment on the CIS and CGAS. Correlations between the
FISCA score on the one hand and scores on the CIS and CGAS on the other hand,
were higher for the FISCA total score than for any of the FISCA subscale scores.

Close examination of correlations with the two global measures of functional

impairment indicate that each of these measures was linked to somewhat different

67

68

aspects of the FISCA, with correlations between the CGAS and FISCA subscales more
often signiﬁcant than correlations between the CIS and the FISCA subscales. One way
of explaining this pattern is that the CIS uses very global and ambiguously anchored
questions about functioning (e.g. , How much trouble does the child have getting along
with his/her mother?), w hereas the FISCA and the CGAS rely more on descriptions of
speciﬁc behaviors (e.g., skips school, hit someone, tried to hurt self, etc.).

More generally, correlations with the CIS and CGAS generally support the
validity of the FISCA (especially total FISCA score) as a measure of functional
impairment. Notably the FISCA’s advantage over both of the other global impairment
measures is that it allows the clinician and researcher not only to measure overall

impairment, but also to pinpoint speciﬁcally where the child's impairment lies.

 

ParenLreport. The FISCA was also hypothesized to be related to parent’s reports of
child symptoms. First, the total FISCA score was hypothesized to correlate with the
total Child Behavior Checklist Score (CBCL). Secondly, each of the FISCA subscales
were hypothesized to correlate with speciﬁc CBCL scales measuring similar or closely
related constructs.

As expected, the total FISCA score correlated signiﬁcantly and positively with
the total CBCL score. The total FISCA scores also correlated with scores for the
CBCL internalizing scale and the CBCL externalizing scale. However, the total FISCA
score was more correlated with the externalizing scale than with the internalizing scale.

The FISCA subscales also were examined in relation to the narrow-band CBCL

scales. These data supported the FISCA’s concurrent validity. Eighteen of 19 a priori

69

predictors were substantiated, i.e., the FISCA subscales and the CBCL scales
measuring conceptually similar problem areas, did in fact appear to assess similar and
related phenomena.

The individual FISCA subscales were also correlated with the CBCL broad-
band scores. In short, the FISCA appears to assess impairments in functioning in
relation to externalizing behaviors more strongly than it does with functioning in
relation to internalizing behaviors. Most of the FISCA individual scores were more
highly correlated with the CBCL externalizing scale than the CBCL internalizing scale.
As would be expected, the only scale that was more correlated with the internalizing
CBCL score was the Moods and Emotions scale.

This may be due in part to construction of the FISCA, in that the only scale that
focuses exclusively on functional impairment as evidenced by internalizing behaviors is
the Moods and Emotions Scale. The Home, Community, Being with Others and
Alcohol and Drug FISCA scales mainly assess impairment in functioning in relation to
externalizing difficulties. In addition, the School and Thinking scales also tend to
emphasize externalizing problems; although internalizing problems’ contributions to
impairment in these areas are tapped as well. On a more theoretical plane, there is a
plethora of evidence in the research literature that externalizing problems have a more
chronic and pathonomic inﬂuence on functioning than internalizing problems (See
Kazdin, 1985).

Nine correlations between the FISCA subscales and the CBCL subscales that
had not been hypothesized were also statistically signiﬁcant. Although these

association (e.g., between home and attention problems and thinking and social

70

problems) made theoretical sense, predicted relationships between the FISCA subscales
and corresponding CBCL scales generally were stronger than signiﬁcant but not
predicted relationships. Notably, 4 of the 9 non-hypothesized correlations were
between various FISCA subscales (e.g., Thinking, Home, Being with Others, and
Mood) and the Attentional Problems scale of the CBCL. Attentional problems
undoubtedly disrupt functioning in many areas of children's lives. In fact the attention
problems scale on the CBCL correlated signiﬁcantly with almost all of the FISCA
subscales with the one exception of the Community subscale. The Attention Problem
Scale correlated most strongly with the Being with Others FISCA subscale and the
Thinking FISCA subscale. This ﬁnding is consistent with the literature that shows that
children with ADI-ID are impulsive, inattentive (which the Thinking scale taps with
items such as, “ﬁnd it difﬁcult to remember things") and often engage in antisocial

behavior (e.g., Hechtman & Weiss, 1983; Ross & Ross, 1982).

 

dim The FISCA was hypothesized to be correlated with child report of behavior
problems and symptomatology. However, lower correlations were expected between
the FISCA and child report of symptoms than with the FISCA and parent report of
symptoms. The importance of informant differences has been discussed elsewhere
(Achenbach et al. , 1987). Generally, the literature indicates that parents seem more
capable of reporting on behavioral manifestations of emotional difﬁculties in their
children but have more difﬁculty identifying the nature and intensity of their child’s
internal feelings (Edlebrock et al., 1986). A similar ﬁnding was reported here. The

School, Home, Community subscales of the FISCA assess overt behavioral difﬁculties

71

(as observed by the parent), and it was these scales that correlated most strongly and
consistently with youth report of problem behaviors. However, ﬁndings for the Being
with Others subscale were inconsistent. This scale highly correlated with parent report
of child aggression and child externalizing behavior but was unrelated to youth report
of aggression and externalizing behavior. This may have to do with differences in child
and parent report of aggression across different settings. The CBCL does not specify
which context the aggression is occurring in, whereas even more importantly, parents
are probably less aware of their children’s aggressive behavior with peers, than their
behaviors at home; this can help to explain why the Being with Others FISCA subscale
and youth report of aggression do not correlate. Overall, the correlations between the
FISCA and the YSR were low and provided relatively weak support for the concurrent
validity of the FISCA. This may be a result of parent-cth informant differences. In
addition, it may be that parent report of functional impairment does not relate to child
report of symptoms and instead relates to other factors (such as family relations or
number of hospilizations).

The FISCA also was expected to correlate with adolescents’ report of their own
alcohol and drug use. Youth reports of alcohol and drug use consistently and positively
correlated with parent report of child substance use on the FISCA (r’s ranged from . 18
to .53). Relationships between the FISCA and the alcohol and drug measures generally
were stronger and more consistent then relationships between the FISCA and the YSR.
These results suggest that the relative inﬂuence of informant differences may depend on
what is being reported. The FISCA alcohol and drug scale asks about not only how

much children drink, but also about whether children are drinking or doing drugs to

72

such an extent that they are having legal problems, are injuring themselves, or are
addicted. It looks at impairment in functioning due to alcohol or drug use rather than
at underlying motivations or emotional difﬁculties that could result in alcohol or drug
use. The youth self-report alcohol and drug measures also asses not just quantity but
also consequences of using alcohol and drugs. Parents may be more aware of child
behavior when it results in social disruption, so while they may not precisely know
what or how much alcohol or drugs their child is using, they will know when it is
disrupting the child’s functioning.

InpatientxersusﬂutpatienLFISCAmm. The FISCA’s discriminant validity
was demonstrated by the ability of the total FISCA score to discriminate between
inpatiernts and outpatients. That the FISCA Home, Self-Harm, and Alcohol and Drug
subscales in particular, discriminated best between patients and outpatients makes a
good deal of theoretical sense; suicidality is a major criteria for admission, also parents
of those who are seveme impaired in the home are more likely to seek an out of home
placement. Similarly, if a child is addicted to alcohol or drugs or is in legal trouble
because of drug or alcohol use, parents are more likely to feel their child needs to be
hospitalized. Inpatient facilities typically use these problems in their admission criteria,
especially when outpatient treatment has failed. An Obvious caution in interpreting
these results is that the FISCA to some extent was used by the participating facilities to
make level of care decisions. However, the FISCA was one among a number of
factors used to make this decision. Other factors, such as available insurance, whether

or not the parent was asking for inpatient care, and history of failure of outpatient

73

efforts played a large role as well. Moreover, the FISCA has only rwently been
introduced to the facilities and during the study period was frequently not scored or
even consulted until after admission. In addition, care was taken to exclude patients
from the study who had been assessed for inpatient care but referred to the outpatient
clinics.

Diagnoses. The FISCA was hypotlnesized to be able to discriminate between
patients with no, low, and high comorbidity, as well as betweern patients with and
without a particular diagnoses. This study showed that the FISCA could in fact
discriminate between patients with different levels of comorbidity. As expected,
individuals with more diagnoses presented as more impaired on the FISCA.

Shaffer et al. (1989) hypothesizes that comorbidity is the rule rather than the
exception in childhood disorders. This study supports his theory. In this study, only 5
out of 75 (6.7%) children had no comorbidity. Most importantly, number of diagnoses
(one, two, more than two) related in a linear fashion to impairment scores on the
FISCA, so that while comorbidity may be the rule, extent of comorbidity may also be
important.

In general, the FISCA also was able to discriminate between children with or
without speciﬁc diagnoses. Many of the ﬁndings were as predicted; i.e., the School,
Home, Community and Being With Others scales discriminated between patients with
and witlnout an externalizing diagnoses; the Alcohol and Drug scale discriminated
betweern patients with and without a Substance Abuse diagnoses; and also the Thinking
scale discriminated between patients with and without a Thought Disorder diagnosis.

Because there were only three individuals with a diagnoses of a thought disorder this

7 4
last result is not very deﬁnitive.

The nnood scale was the one subscale that did not discriminate as predicted, i.e. ,
between the presence and absence of an internalizing diagnosis. However only 7 of the
58 children receiving an internalizing diagnoses did not also have an externalizing
diagnoses. Some theorists argue that the negative social consequences of aggression
cause depression (Patterson & Stoolnniller, 1991). For example, there is evidence
(Pope, Bierman, & Mumma, 1991) suggesting that one consequence of hyperactivity
may be per rejection, which in turn may lead to internalizing problems (Hymel,
Rubin, Rowden, & LeMare, 1990). Therefore, one reason the Mood and Emotions
Scale may not have been able to discriminate between patients with internalizing
disorders and patients without internalizing disorders is that the internalizing diagnoses
group in this study was not accurately assessing ”pure” intemalizers. Many children
with internalizing diagnoses may have been experiencing the secondary effects of
externalizing problems and even those qualifying only for an externalizing diagnosis
may have suffered from depressed or anxious mood.

The inability of the Moods and Emotions scale to discriminate between the
internalizing (and not-internalizing) groups also may be related to informant issues.
The FISCA is completed by the parent only, whereas the diagnosis is given based on
interviews with botln parent and child. If a child denies being depressed across the
board, the child would not receive a diagnosis of depression (even if the parent reports
depression). Therefore, there may be children whose parent report would lead to a
diagnosis of depression (and would show mood problems on the FISCA), however the

child would not have received a diagnoses of depression because the child denied being

7 5
depressed.

In general, the FISCA scales were relatively successful at classifying cases into
diagnostic groups. However, the results also support the hypothesis that functional
impairment and diagnoses are not identical. Often, it was not a particular scale, but
rather a combination of scales that was able to discriminate so that multiple
impairments were likely to be related to diagnostic groups.

IengthnLStay. As expected, the FISCA score at the initial assessment was
able to predict length of stay, substantiating the predictive validity of the FISCA.
However, this relationship was stronger for the Nebraska facility than it was for the
Michigan facility. This may be due to the fact that the Michigan health care market is
almost twice as penetrated by managed care organizations than the health care market
in Nebraska (19% versus 9.9%; HMO-PPO Digest, 1995). The way in which
managed health care views psychological treatment has begun to affect the ways in
which hospitals work with their patients. Managed care has put less emphasis on initial
level or changes in functioning, but instead has examined the patiernts on social support
arnd sought to decrease acuity oftlne problems so that the child can be moved the child
to a lower level of care. More recently, the aim of inpatient treatment has changed
from 'a total reconstruction of the personality", to stabilizing the patient so that they
can receive less restrictive and less expensive treatment (Nurcombe, 1988). These sorts
of changes in health care policy are likely to weaken the relationship between
functional impairment and length of stay. The fact that external factors such as

insurarnce company policies affect length of treatment has important implications for

76

botln research and clinical work. Researchers must be aware of this confounding factor
and clinicians must assess the effects that external demands (such as managed care) are
having on the treatment goals as well as on the quality ofcare their patients receive.
EntureDirections

In sum, this study supports the validity of the FISCA as a measure of functional
impairment. Three main theoretical questions need further exploration. The ﬁrst
question is how much parents and childrern agree on functional impairment. Studies are
needed to unravel the importance of differences in child and parent report of the child’s
functional impairment. It has been shown that both children’s and parents’ reports of
child behavior are valid, but that they relate to different things (Kazdin, 1985). Future
studies need to focus on ways in which this is true, for example how child and parent
report of child functioning relate to number of hospitalizations, number of criminal
offenses, etc.

Consistent with previous research, this study shows that parents and children
agree more on externalizing difﬁculties than internalizing difﬁculties (i.e. , child report
on externalizing problems correlated with the FISCA subscales, whereas child report of
internalizing problems did not correlate with FISCA subscales). This study was not
able to directly assess if parents and children agree on functional impairment, since
children did not complete a functional impairment measure. In addition, because the
data came mostly from motlners, this study could not assess whether or not the
relationship of the person who reported (i.e., mother, father, etc.) affected the
relationship betweern parent-child agreement. Future studies assessing the relationship

between child report on the FISCA (a youth form was recently developed) and parent

77

reportonaFISCAwithalargesampleofbothmothersandfathersareneededto
address this issue.

A second tlneoretical question has to do with the interrelationships between
functiornal impairment arnd symptomatology and functional impairment and diagnoses.
It is important in light of the original assumptions with which this study began, to
realize that, while the FISCA does indeed relate to measures of parent
symptomatology, none of the correlations between the FISCA scales and the CBCL
were greater than .52. Hence, it would appear that the FISCA is indeed measuring
something overlapping with but also conceptually distinct from symptomatology. The
relationship between diagnoses and functional impairment is less clear because of the
smaller sample size used to assess these associations. For instance, only three
individuals had a thought disorder which made it impossible to reliably assess the
relationship between the FISCA thinking scale and diagnoses. Membership in the
externalizing, internalizing and substance abuse categories were larger, making those
results more reliable.

A second difﬁcultly in drawing conclusions from the results for the diagnostic
sample is that approximately half of the patients had both internalizing and
externalizing diagnoses (and most intemalizers were also externalizers). This made it
impoka to assess the distinctions between 'pure' internalizes verses ”pure”
externalizes in functional impairment. Future studies using a larger sample are needed
to examine these questions. It is also important to note that the relationships between
the FISCA and symptomatology, and the FISCA and diagnoses were only tested and

supported in the inpatient sample. Symptom data and diagnoses for the outpatient

7 8
sample was not available. In addition, relationships betweern diagnoses and symptoms
couldnotbe assessedintlnisstudybecauselesstlnanhalfofthepatientsintlne
diagnostic sample also had parent or child symptom report data.

The ﬁnal question is how the FISCA can be used by clinicians. The FISCA’s
ability to discriminate betweern inpatients and outpatients suggests that the FISCA could
be very useful to hospital admission and intake staff. The FISCA could be used to
effectively and quickly decide the appropriate level of care for patients. Studies are
needed to assess what score on the FISCA indicates a need for inpatient treatment.
Future studies could assess wether or not the FISCA is more useful than diagnoses or
symptom measures in making level of care decisions.

In addition, this study suggests that the FISCA could be used to make effective
treatment plans. Currently, the emphasis in inpatient treatment is on problem focused
treatments. The FISCA could be used by therapists to quickly identify the problem
area. This information could be used at face value and the therapist could design a
treatment plan to lower the child’s impairment level. In addition, it might be useful for
tlnerapist to look at contrasts in child’s functioning. For instance, if a child is impaired
at home and not school, this would indicate that there is something about the home
setting and/or the family that is contributing to the child’s difﬁculties. As a result, it
would be important to include changing maladaptive family patterns into the treatment
plan, if long-term change is desired. However, if the child was impaired in school and
not home, goals of how to improve child’s functioning at school would be included in
the treatment and therapists would want to consult with the child’s teachers. The

FISCA does look at the context of the problems (e.g., home, school, community),

79

whereas diagnoses and symptom checklists do not. This suggests that the FISCA might
be more useful in treatment planning than symptom checklists or diagnoses, in that the
FISCA allows the clinician to learn not only about the child but also about the child’s
environment.
Methodologicallmblems

Several other metlnological problems need to be addressed. The sample
consisted mainly of Caucasian children. This limits the generalizability of the ﬁndings.
A study using a more racially diverse sample is needed to assess if the FISCA is a valid
instrument of functional impairment in different ethnic groups. The study also lacked a
control group of ”normal“ children in the community. This study was not able to show
that the FISCA can discriminate between "normal" children (non-clinical sarnples) and
tlnose receiving outpatient services. An analysis of these differences would be
confounded by the fact that not all children who need psychological services actually
receive them. Nevertheless, one would expect less impairment in a normal (i.e., non-
clinical) group-
Conclusions

While this study did have several ﬂaws, it is an important ﬁrst step in validating
the Functional Impairment Scale for Children and Adolescents (FISCA). This study
shows that the FISCA scales are related to both symptoms and diagnoses, but are not
merely measuring the same construct. This study helps clarify the differences between
functional impairment on the one hand and symptomatology and diagnoses on the other
hand. Functional impairment, as measured by the FISCA, focuses not on symptoms per

se, but rather on what areas of the child's life are disrupted by symptoms. These data _

8 0
show that a child with a particular diagnosis may be impaired in several different areas
offunctiorning sothatthereisnotonedirectlinkbetweenanareaofimpairmentanda
particular diagnosis. For instance, a child with an externalizing diagnoses is likely to
be impaired in school as well as his or her relation to peers and also may (or may not)
have impairment in thinking associated with ADI-ID.

This study is unique in that it looked at a multi-dimensional measure of
functional impairment. Most existing measures of functional impairment assess global
impairment rather than speciﬁc areas of impairment. Global impairment does not
allow the clinician or the researcher to focus on the precise areas in which the child's
difﬁculties lie. This study showed that the FISCA is able to accurately assess
impairment in functioning in a variety of areas. Its advantage over more global
measures make the FISCA a better instrument for treatment planning and level of care
decisions, as well as for evaluating speciﬁc treatment outcomes. In conclusion, this
study shows not only that the FISCA has concurrent, discriminant, and predictive
validity but also that measures of global impairment, such as the FISCA, are an
important and needed addition to traditional assessment protocols that focus only on
symptoms and diagnoses in assessing psychological difﬁculties in children and

adolescents.

APPENDICES

APPENDIX A

81

APPENDIX A

 

 

 

 

 

 

 

 

 

EISC! C 0' 0
SCHOOL
Seesaw Medal-aw Whpdr-d
scnoor. ”meshed“ 01 Wench-ed N Mikes”. 11
efechoelrd-lerphehh sogolosdoequroxnly mire-school
(huh-«worried» oeceeverytwowedsor-ore) (oeceano‘orleee)
gobscheoo
Pound-km Q Perforn'qhelovabiiw 12
Who-school 02 gredepoin'ubelowevenge. hschoolbeculeoffaihre
“mutant“ bouquerH'.
We‘dscheel 03 eceddcceenes Worm
dletvork'ng orelcppywosk
Marne-cow“. 10
”Mum 0‘ Warm WM-03 13
resuled'np-iiveectieesor hehviorresuhed'as regriredcleeeslpervnsee
poorecede-icperfor-ce Wad/erectin- lyteecherbccerleof
whacked-Iberian “View
Pdl'nalaadedccbu; 8 Woe
uﬂelysogoeetoheeext WM
grade
Mint-HM“ 14
Mir-slaw “ VM
orpeerorpﬂoﬁersdrﬂ
bybs'lg'ngaweqsetescheol
”vior'nsoanefcedol fl
“aim
age-Nahum
HOME
Sev-el-fﬁ-d Mel-w Mill-1dr“
nos: R-myﬁonnhesee 15 Ramiro-shone 18 Prequedyreﬁrseseodo 21
Manchu-ore w cliche-1y." age-Wm;
dwm Wmhowetoﬁe ”Merely-remade“
”Worn-rd“- pares. delimit-charm
uddudsforrnse-He
Mar-sown l6 Push-”Mum 19 behevierhtehe-e
«moths/ﬁeuto “mu-forage
hemedﬁouﬁehe-eor witchehviorhtlnehene,
.yseeehvehbere-everl chronncallyfailedtoneaage
fro-khan qnproprisnexpechtiou
“summit
Cth-eelyhe 17 endurancemrdem
Wishbone
“MM Repe‘edlyfeiledso 2O

 

 

 

 

 

 

 

82

 

 

 

 

 

 

 

 

 

COMMUNITY
Snow Medea-dew my“
CW] Mdeeveseb 22 “ﬁend/aw 26 Denied-diner 29
[MAI- hgedpropeseyoﬁde Mafﬁehm-wlvig Vic‘s-dish
bins Mollie-whavicthhg.
mdopﬂﬁ'ng. v-rlel'nl.
“bu-him“ 23 ochoductle-‘l'ngtoa
MM“ 24
cr‘lbﬂvierm ”Motif-e 27
Morin-haw
ersevuehvioﬂieek...“ (hMaderceert 28
MM.m-pﬂlﬂ W
dabbdea'hgorc-ryig
~.Mwﬁem
N—‘uaﬂysical‘.
“annexe-lad)
Dehrpedorcr'ls'nel 25
behavioral-“hm
«Misha-er
dado-o‘er
THINKING
Berni-pd!“ Modes-dew WW
I“ C-etbe'naeornlschool mathehviornsore Mamier-ese
charm-amused Mathilda-ulnar hhs'reor'qaiedlmoler
Honor—1W0. .eagechilrlrunANDdliH meagechﬂrubeceueofm
“MMH reg'mesaqnec'ulechoolprogrll “Item:
Osman-yo! orqec'dsnperv'liesnhecesneof
hem: mofﬁefolcw'lg:
Oil-severelyd'ncnmdu MdemM Inert-helm, 39
teheerpheewrm mm,w Worm
Marnie-calms MMwiI-ges. M.“
celeb-linemen repetiionefwonlsorphreses
Sevese'qair-sn'n 31 ofwoshorphr-es
nﬂyu’um- Whack“ 0
“cue-“aura Athed'uos'iﬂed.“ 35 “(hubbbem
'b'lm're'behvier m rm-m-u- per-noid.h-oddbelieﬁer
Quayle- “WW,
Cali-ionisogruhn Ifolderlhan8yeers.u¢egesh
“boll-med B'nrsepseocchensor 36 ngnalth'esk'ngMievesheor
heme! ”‘1th shecurneketh'ngshqpenpn
Whethdeviord oroccehin-a ”mm-nun)
robbeheviorer
Wuhan Realiytesti'inaiedby 3‘1 Childocc-ie-ﬂybeconesso
areeeriosﬂy'lqaled akin-(e.g. MW 41
h-gicalpewere)or ceaﬁnedheordeh-dsﬂiculy
B'narrecem-scdie- 33 mm WWW
“mime-Near
ext-elytho Child‘sco-ﬁnis-eh 30 Speechiowasioaalyodd 42
mam-om herferesw'ehh'nlher «idiosyncratic
”albumen.“ Wendi-Italics:
(e.g. Wm, “M“
ecbhh, id'soey-crnic hilherbdnvior

 

 

 

 

 

 

 

BEING WITH OTHERS

83

 

 

 

 

 

 

 

 

 

lee-ohm Medea-dew hill-pie“
mom cease-aw Persud'ﬁiculy'nhehg 41 Tubbhvedﬂiculy 52
arms frieﬁ 43 Hedorget'lgaleq'hpeese herect'ngvhhpeesshecmeef
”Mi” “(imam Me
”Warm hehvbr
Sunﬁsh-sewing”
Maryanne 44 Meredith—b O “awn-deﬁes!
nhh
Wipes” 0 WM” 34
Prepddleresrie- ‘ m.m¢u behest-germane
WWW“ Methane...“ WWW-ca
“evens-Woke. “new
mas/«mun d-gaheleag'lgsefoﬁere. Mature 53
Mummies-bee Herbalism) “mepﬂfeﬂ
mwmdc‘ﬂu bphy'hye-gerchiﬂsu
whine- r-ovedfle- MW“ S
heneerechml nil/orqerceﬁol
Winch-ouster 4‘ WWW 51
“mufﬂe-es Meters
oreppesheeea
MOODS 8: EMOTIONS
W Moedpseblss-le 56 Perehesnteuotieasl 60 PerledicM'nnood 65
m swan-HUM“ dyereguﬂien;nrkedlnood ordifﬁcehiseexprus'ug
swings,bhlstedaffect. anodes!
Extras-sod“ 3'7 cruiser-candied
' ' expre-iveee- Periodicalyexperiaces 66
U-nlorveryiue sad-QJMMM
eapsessseeofmotiouﬂotere Severeequisoenxiuy; 61
see-chlorm ems-yrequieqnec'nl Overlysel-crﬁcel. 67
w(e.g.sleqf-gwﬁ perfecﬁeehieorsdveto
Worm-e ﬂ ”)tonr'nhized'dn- aiichhyotese

 

 

 

Pere'deudepruionor 62
‘erlon'nﬂissofple-src)
aecompnniedlryseeﬁc
orsleqidifﬁculies
Pere'utsdnxietyaeconpuiedﬂ
hyeouicconphhs. poor
Mahdifﬁculies
MW ‘4
Mof mx’norbul'en'n

(e.g. weigh Ion. vol-ﬂing.
“Isis Ii mis- '0‘

 

 

 

 

 

84

 

 

 

 

 

 

 

 

 

 

SELF-HARM
laurel-ﬁnd “new h-dhpdsud
sun'- MI-Udtokilerht 70 “Matinee“ 71 Ween-ace“ 72
HARM Whammy» HIM-away“ mmw—m
rﬂ “self—hum; udcncies,buverynnﬁelyte
h serio- eel-ﬁsry orde‘ hehdes “ﬁfe-M cue seriou- hjlssy, e.g.,
Harland humeuivislgesesresuuicirlel wmorecrdchhg
Mummu tethwhaﬂohject
hdmorsdf-mub
”bulimia.“
«ﬁt-knapmer
c‘hrneseelr
ALCOHOL & DRUG USE
Scream Mill‘s-d “hm
Workiﬁﬂ H'ghor'ﬂoxicadoece 78 Reguhrue(nledaee 83
(Al-COBOLI (uh-strobe“) em ewumm
uncle: llqeﬂﬂ'w- 7’ «been-ﬁght
ﬂuency-per.“ 74 Maugham“
Madmuceeweskor ofﬂine Hml2eryeqc,ues 84
more M(e.g.afew

Alcoholordregnleresuhed ﬂ these-0‘)
Severephyeicalorueﬁeeal 75 hipsiressd'nrolef-cu'onhg
WCWNW theme'nschoolormiy Occ-le-ﬂyhoxicdeddnseﬁ

(e.g. physicalcrev'ngs.uee'n (e.g.doeun'tdochores,breaks unborn-yearns-
mhg.hvﬂreval mule-nachoolorvork. W
ml Machetes-hear“.

whichever” pickedqbytepolice,vioha

W) curfew

Ueeefaleehelerdregs 76 Alcehelerdrlg-elerlto 81

Whmerely'qeiedrole eegdveeoc'nlcosnerpsaceﬂeg.
Mamba.“ codlichwithflnilyorfrids.

“cremation.“ofjoh. soc-Imam“

“accusation-heat “bee-embers)

Mme-big

afeleay.or'.5pigc§ers) Alcoholordmg-ernledhﬂ
muwwmmu

M'IIeofalcoblordrlgs 77 ri(e.g.driv'nglllerte
“mp‘feh‘r'lk whence,
chiklwashkeadmgeof
sexully,hel¢eexperiaced
lhorhea‘probhl)

 

 

 

 

 

 

APPENDIX B

85 85

 

Instructions: Duringthiaassessmerntwearegoingtoaskyouanumberofquestionsaboutthingsthatrrnayhave
happenedduringthepastthreemonths. Thetimelinewillhelpyoulocatethepastthreemonthsinyourmindby
helpingyoutorememberimportantdningsthethappenedduringﬂnatﬁme. Usethefollowingstepstoﬁllinthetime
line.

1. Writeintoday'sdateonthelinethatsays Today.

2. Wrieinthedatesforeachoftlnepsstthreemonthsonthelinesbelowthe

3. Have any major holidays occurred in thepastthreemonths? Write the
holidaysonthe

4. Has your family had any birthdays intlnepasttlnree months? Writein any
birthdaysonthetimeline.

5. Have any important things happened in your lifeinthepasttlnreemonths?

Writeanyimportantthingsduthappenedinyourlifeonthetimeline.

 

 

3morlhsago 2Mago lmorehago Todayis:

 

APPENDIX C

 

WeanukingywbpasﬁciptemasnslytowahuemusasmeuinunmeucdkdtheChild
BehvioeringForm(CBRF)thatweaseusingatRivmllofMichiganHospital.Weareusingthis
imtnnneu each day to miter the negative and positive behaviors of chilrl and adolescerl patients admitted
toourfacil'ty.Thepresertstudylooksatthewaysthatbehaviorsexhibitedbychildrenandpatierlsrhtring
Mrbspidmymluewdhmudxdmduthethmtheymdmiuedmthebsphd.hformuimm
ywproviderhuingaqueddiagmsﬁcirnrviewwiﬂbeusedfmthisshxly,andwillalsobegiventothe
meddheahhbambusiumdaqmimngthemuappropriuenumapprowhforymchiﬂ.

Pleasecarefully reviewtheinformuionbelow.Ifyouhaveanyrpestionsatall,besuretoaskthe
pamwhoisreqwsﬁngymrpaﬂicipﬂionwdﬂywﬁﬂyundashnddnpmcednumvohed.Ifyoudo
agreetoparticipate,uatokenofmnappreciation,ywedﬂreceiveagiﬁof810aﬁerﬂsetelqnlnne
irierviewiscompletedmvenifyourchilddeclinestoparticipate.

I, ,parerland/orlegalguardianof ,aminor
child who rs curremly received psychiatric services at Riversiell of Michigan Hosp'nal, have been fully
mammmmdmmmmmrmmmmipm. I
havebeen informed of andunderstandeachof the followingpoirns(anyreferencesbelowto'my child“ refer
totheabovenamedchild):

1. BysigningthisagreemulJamgivmgpermissionfmauaﬂmenbaﬁomRivendeﬂm
talqilmemewithinthenextﬂharrswiththeimerlofirlerviewingmeehoutmychild'semotionaland
behavioral ﬁmctioningpriortoerleringtreatmerlatRivendell of Michigan Hospital.Iamalsogiving
permissionforastaﬁ'menlnerfromRivendelltointerviewmychiklwithin48hoursofcompletingthe
irlerviewwithme. TheiderviewswiﬂuchukeapproximnelylhmrwcompldeJunderetamlthuthe
procurtoubctparalsmbecdledismndom,andthulmaynotneceswﬂybecaﬂed.

2.1Mﬂrweivea8mgiﬂnfqmchtkmfmparﬁcipaﬁngmthesndymmewkphom
inerviewiscompleted.Iwillreceivethisgiﬂevenifmychilddeclinestoparticipateintheirlerviewor
doesnotconpletetheirlerviewwhenapproachedbytheinterviewer.

3. Ihavetherigllwhencalledtodecidenottoparticipateinthestudy,ortodeclinetomswerany
speciﬁc rpestion(s) that I am asked, or to discomiraie my participation at any poirl inthe islerview. Ihave
therigll,regardlessofwhetherornotldecidetoparticipate,todeclinetopermitmychildtoparticipate
shouldheorshebeaskedtodoso.Mychildalsohastheriglntodecidenottoparticipateinthestudyﬁo
declinetonswer speciﬁc qrestims, ortodiscominueparticipationatany poim.

4. Theimerviewwithmyselfand/ormychiklmaybeaudiotapedinordertoestablishimerrater
reliabiliyindeterminingmychikl'sdiagnosis.Nooneotherthanmembersoftheresearchteamwillhave
accesstotheeetapes.ThetapeswiﬂbedestmyednohterthanSmomhsaﬂertheinterviewshavebeen
completed.Intheimerim,thetapeswillbekeptinalockedﬁlecabinetinthehospital'sresearchoﬂice.

5. Anyinformationlormychildprovideintheinterviewswillbetransformedimorarmbersand
«landmacomputerdatafile.ThewrittenrecordofthediagnosticirlerviewwillthenbedestroyedJ
mwmmmmnwummmmmwmmmymmmrm
mychildmigllpmvklethumggeststhumychiﬂmanotherminorisbeingmbjeaedtoorlnsbeen
subjectedtophysical or sexual abuse.

6. Iundernandmmifmychiﬂparﬁcipammthemﬂy,mpraeahkmhacmﬂhsﬁdky,lwiﬂ
mbe'mformedofanyinformationthatheorsheprovides,inchrdinginformationahorlanytransgressions
orillegalbelnviorsthatmigllbeunknowntomeunless,inthejudgmentofhospitalstaﬂ'designatedto
review this information, my child clearly and with minimal doubt is in imminent and serious danger of
horminghim/herselforothere. Ifthissihrationshankloccnn,dnnahospitdstaﬁmemberwiﬂinformmy
M'sdmepiﬂofﬂresecommaﬁﬁeﬂnrapiﬂedﬂmdwﬁhmandmycﬂﬂmdimsdnm
fortheseconcerns.

7.IfIdecidenottoparticipateordeclinetopermitmychildtoparticipate,ormychilddeclinesto

 

87

particirue,hmymaﬂaspecuofthhsmdy,mydechionwiﬂmmmyaﬂeathepossbﬂiqurdityof
caredRiveulellHospial.
8.HIhavemqusﬁomadenshﬂyorIwuikllikemreceivemypubhcaﬁom
thatrn’ghfollowfromtln'sdudchancoslactImrieVanEgerenorDr.SusanFrarkatRivendellHospital
bycall'mgSl7-224-ll770rhywriingto 101 W. Townend Road, St. Johns, MI48879.

 

 

PareslorLegalGuardian Date

APPENDIX D

 

(To be reviewed with minor patients who are at least 8 years of age and are deemed
competent of voluntary consent.)

1, ,have been told all about the study
andlunderstandwhatlwillbeaskedtodo. Iunderstarnd thatthepurposeofthe study
is to see how well the hospital is doing in helping children and families. I also
understand that:

1. By signing this agreement, I am giving permission for someone from Rivendell to
talktomeadayortwoaftertalkingtomyparentsorguardian on thephoneaboutmy
feelings and behavior before coming to the hospital.

2. I have the right to decide not to talk to the interviewer, or not to answer a question
that I am asked, or to stop answering questions at any time during the interview. I
know the interview will take approximately 1 hour to complete.

3. I also understand that the interviewer will not talk to anyone outside the hospital
about what I say during the interview unless staff at the hospital are quite certain that I
am in danger of seriously hurting myself or others, or I or others with whom I am
connected are in danger of serious harm by others. If this happens, then the hospital
staff will inform my therapist, parent or guardian, or other appropriate authorities that
they are worried about me. My therapist will then meet with me and my parents to
explain why the hospital is concerned and will try to help me and my family get some
help.

4. HI decide not to speak to the interviewer, my family and I can still get help from
Riverndell Hospital and no one will hold it against us.

5. If I have any questions or concerns about the study or I would like to receive
information about the ﬁndings from this study, I can contact Laurie Van Egeren or Dr.
Susan Frank at Rivendell Hospital by calling 517-224-1177 or by writing to 101 W.
Townsend Road, St. Johns, MI 48879.

 

 

Child/ Adolescent signature Date

REFERENCES

References

Achenbach, T. M. (1983). Manual_for_the_ChilrLBehaiLior_Checklistand
ReidsedﬁhildﬁehasLiOLPmﬁle. Burlington, VT: University of Vermont, Department

of Psychiatry.

Achenbach, T.M. (1987). ManualionthelonthjelszeporLandhoﬁle.
Burlington, VT: University of Vermont, Department of Psychiatry.

Achenbach, T..M (1991a). MannaLfOLtheChildBehayiOLChecklisLand
Renaisedﬁhrldkhmonﬁmﬁle. Burlington, VT. University of Vermont, Department
of Psychiatry.

Achenbach, T.M. (1991b). Mamralfor_the_Youth_Self;Re;xnrt_and_199_1
Proﬁle. Burlington, VT: University of Vermont, Department of Psychiatry.

Achenbach, T. M. & Edelbrock C. S. (1983). Mannaliouheﬂhildﬁehaidor
ChecklisLandRexisedLhildBehanLiOLPmﬁle. Burlington: University of Vermont,

Department of Psychiatry.

Achenbach, T. M. ,McConaughy, S. H. ,& Howell, C. T. (1987).
Child/Adolescent behavioral and emotional problems. Implications of cross-informant

correlations for situational speciﬁcity. PsychologicaljulleﬁmJﬂl, 213-232.

American Psychiatric Association (1985). DiagnosticandstatistimLmamralnf
mentaLdisorders (3rd ed.). Washington, D.C.:Author.

American Psychiatric Association (1994). DiagnosticmanstatisticaLmanuaLof
mentaLdisorders (4th ed.). Washington, D.C.:Author.

Angold, A. Weissnnan, M. ,Merkangas, K. (1987). Parent and child reports of
depressive symtpoms in children at low and high risk of depression. Joumalnfﬂhild

EsyhmlogyandlsychiatrszAlliedDisciplines,18, 901-915.

Apter, A., Orvaschel, H. ,,Laseg M. ,Moses, T. & Tyano, S. (1989).
Psychometric properties of the K-SADS-P rn an Israeli adolescent inpatient population.
. - : 61-65.

 

89

90

Bird, 11., Canino, G., Rubio-Stipec, M. & Ribera, J.C.(1987). Further
measures of the psychometrric properties of the Children's Global Assessment Scale
(CGAS). AmhixesnfﬁeneraLPsychiatryA, 821-824.

Bird, H.R., Shaffer, D., Fisher, P., Gould, M.S., Staghezza, B., Chen, I.Y.,
& Hoven, C. (1993). The Columbia Impairment Scale (CIS): Pilot ﬁndings on a
measure of global impairment for childrern and adolescent's. InterrnationaLloumaLof

MethndunksychratneResearchJ, 167-176.

Bird, H..R ,..,Yager,TJ Staghezza, B., Gould, M..,S Canino, G., &
Rubio-Stipec, M. (1990). Impairment m the epidemiological measurement of childhood

psychopathology rn the community. ImrmaLoftheAmericamAcademjLoLChildand
AW, 796-803.

Browning, D. L. (1986). Psychiatric ward behavior and length of stay in
adolescent and young adult inpatients: A developmental approach to prediction.

ImnmaLoLConsulnngandﬁhmcalPsychologyJA, 227-230.

Chambers, N.J., Puig-Antich, J., Hirsch, M., Paez, P., Ambrosini, P.,
Tabrizzi, M.A., Davies, M. (1985). The assessment of affective disorders in children
and adolescents by semistructured interview: Test-retest reliability of the Schedule for
Affective Disorders and Schizophrenia for School-Age Children, Present Episode.

AmhmoﬂhneraLEsychnatanZ, 696-702.

Conners, CK. (1989). MmualfOLCmnemLRaﬁngaScales. Multi-Healtln
Systems, Inc. Toronto, Ontario.

Costello, E. J. Edelbrock, C. S. & Costello, A. J. (1984). Validity of the NIMH
Diagnostic Interview Schedule for Children: a comparison between psychaitric and

pediatric referalls.IournaLOf_AhnnrmaLChild_PsycholOgy,_16, 219-231.

Cronbach, L. J., && Meelnl, P. E. (1967). Construct validity 1n psychological

tests. InW.A. Mehrens&R..L Ebel(eds.),Priniciples_of_educationaLand
psycholopcaLmeasurement, (pp. 243-270). Chichago, IL: Rand McNally &
Company.

Division of Mental Health, Mental Retardation & Substance Abuse Services of
North Caolina (1989). NorthLamlinaAdulLEunctionalAssessmentjcale.
Unpublished instrument.

Dunn, O. (1961). Multiple comparisons among means. Journal of the
American Statistical Association, 56, 52-64.

Edelbrock, C. , & Costello, A.J. (1988). Convergence between statistically
derived behavior problem syndromes and child psychiatric diagnoses. JonmaLof

91

Ahnrmnalﬁhildlsychnlogyrlﬁ, 219-231.

Edelbrock, C., Costello, A.J., Dulcan. M.K., Kalas, R., & Conover,
N.C.(1985). Age Differences in the Reliability of the Psychiatric Interview of the
Child. Child.Deuelnpment,.56, 265-275.

Edelbrock, Costello, A., Dulcan, M. K. ,,Conover N. C. & Kalas, R. (1986).
Parent-child agreement on child psychiatric symptoms assessed via strucutred

interview. IonmaLnfChrldPsychratrandPsychologle, 181- 190.

Edelbrock, C., & Achenbach, T. M. (1980). A typlogy of Child Beahivor
Proﬁle patterns. Distribution arnd correlates for disturbed children aged 6—16. Iournal

WW 441-470.

Endictott, J, Spitzer, R..,L Fleiss, J.L., &Cohen, J. (1976). TheGlobal
Assessment Scale. WMM6—771.

Fergusson, D. M., & Horwood, L. J. (1987). The trait and method components
of ratings of conduct disorder-Part II. Factors related to the trait component of conduct

disorder scores. IonmalnLChrlrLPsycholngyandPsychratryJK 261-272.

Forehand, R., Lautenschlager, G.J., Faust, J. & Graziano, W.G. (1986).
Parent perceptions and parent-child interactions in clinic-referred children: A
preliminary investigation of the effects of maternal depressive moods. Behanzionrr
BMW, 161-172.

Frank, S. &Burke, L. (1992). DendealrazatrorLaanautonomanlate
- . : - ~ » ~ ~ - ' 1‘ Paperpresentedattlne
Bi-annual meeting of the Soicety fOr Research on Adolescence, Washington, DC.

 

Frank, S., & Paul, J. (1995). The Functional Impairment Scale for Children
and Adomcents. Unpbulislned manuscript.

Frank, S. &Poorman, M. (1993). Ihcparentnngallranceasamnderatonof

 

at the Society for Research in Child Developemnt, New Orleans, LA.

Gleser, G.C, Seligman, R., Winget, C., & Rauh, LL. (1977). Adolescents
view their mental health. loumaLoflouthandAdolescenmh, 249-263.

Gordon, R. E. ,Jardiolin, P. & Gordon, K. K. (1985). Predicting length of
hospital stay of psychiatric patients. AmenmnlnumaLoLPsychrmﬂ, 235-237.

Gordon, R.E., Vijay, J., Sloate, S.G., Burket, R.,& Gordon, K.K. (1985).
Aggravating stress and functional level as predictors of length of psychiatric

92
hospitilization. HospitalandCommunityPsychiatryJﬂl), 773-774.

Gordon, R. E., Plutzky, M., Gordon, K. K, & Guerra, M. (1988). Using the
Axis V scale to evaluate therapeutic outcome of psychiatric treatment. Canadian

Immalnfﬁsychrm, 194-196.

Green, R. S., & Gracely E. J. (1987). Selecting a rating scale for evaluating

services to the chronically mentally ill. Community_Mental_Hsalth_1oumal,_23_(2), 91-
102.

Green, B., Shirk, S., Hanze, D. &,W,anstrath J. (1994). TheChildren's
Global Assessment Scale in Clinical Practice: An Empirical Evaluation. IoumaLnfthe

WW1 158-1164.

Hammen, C. (1988). A longitudinal test of the attributional vulnerability model

in children at risk for depression. BritisthnmalnLClinimLPsycholngy,_27_(l), 37-
45

Haley, 6., Fine, 8., Marriage, K, Moretti, M. ,& Freeman, R. (1985).
Cognitive bias and depression in psychiatrically disturbed children and adolescents.

WWW 535-537.

Hechtman ,L., Weiss, G., Perlman, T., Hopkins, 1., & Werner, A. (1981).
Hyperactivity as young adults: Prospective ten-year follow-up. In K. D. Gadow &

 

Westview Press.

HMO-PPO Digest: Managed Care Digest Series. Kansas City: Hoechst Marion
Russell Inc. (1995).

Hodges, K. ,,Cools J., & McKnew, D. (1989). Test-retest reliability of a
clinical research interview for children: The Child Assessment Schedule (CAS).
. . , 1 317-322.

 

Hodges, K., Gordon, Y., & Lennon, M. (1990). Parent-child agreement on
symtpoms assessed via a clinical research interview for children: The Child Assessment

Schedule (CAS). Inumal.of.Child.Psycholngy.aanEsychiatry,.31, 427-436.

Hodges, K. ,Bickman, L. ,& Kurtz, S. (1991). Multidimensional measure of
level of functioning for children and adolescents. In_A_Algarin_&_R.M_Er:iedeman

 

chrldren.s.menatlo.helath.(pp_l_49_154). Tampa, Florida: Research and Training
Center for Children' 3 Mental Health, University of South Florida.

93

Hodges, K. (1990). MamnaliontheﬁhildandAdnlescenLEunctional
AssessmenLSrale. Unplublished manuscript.

Hodges, K., Kline, J., Stems. L, Cytryn, L. & McKnew, D. (1982). The
development of a child assessment interview for research arnd clinical use. IournaLOf

AhnnnnathildﬂychologyAﬂ, 173-189.

Hymrel, S., Rubin, K.H., Rowden, L,, & LeMare, L., (1990). Children’s
peer relationships: Longitudinal precieion of internalizing and externalizing problems
from middle to late childhood._ChilrLDevclopment,_61, 2004-2021.

Ihilevich, D, & Gleser, G. (1982). Evaluatingmentalhmlthpmgramszlfhe
ngmssEvaluationSclaes. Lexington, MA. D. C. Heath.

Iverns, C., & Rehm, L. (1988). Assessment of childhood depression:
Corresorndencebetween reportsbychild, mother, and father. ImmalaﬂtheAmerican

Academy.nf.Child.and.AdnlescentPsychiatry,lS,.91-98.

Jensen, P. S. ,Traylor, J., Xenakis, S. & Davis, H. (1989). Child
psychopathology and rating scales and interrater agreement; I. Parents’ gender and
psychiatric symptoms. ImnmalnﬂtheAmerimAcadmynfﬂhinaanAdolescent
W442-450.

Kazdin, A. & Esveldt-Dawson, K. (1983). Assessment of childhood
depression. WWW 157-164.

Kazdin, A. (1985). Treatment of Antisocial Behavior in Children and
Adolescents. Homewood, IL: Dorsey Press.

Kazdin, A., Moser, J., Colbus, D. & Bell, R. (1985). Depressive symptoms
among physically abused and psychiatrically disturbed children._loumal_of_Ahnormal
Psychology,_94, 298-307

Keraus, J. W. (1981). EvaluatiorLoﬂaﬁlohalAssessmenLScaleIOLChildrenjn
anmtmtieanetting. Unpublished Master' s Thesis. University of Arkansas,
Fayetteville,AK

Kolko, D. J. & Kazdin, A. E. (1993). Emotional/Behavioral problems in clinic
and nornclinic children: Correspondence among child, parent and teacher reports.

ImmaLnLChiHEsychologyandEsychiatry,34_(6),-991-1006.

Liebowitz,JH., Rembar,..,JC Kernberg,P..,F Frankel AK &Kruger RS.
(1988). Judging mental health-sicknesss rn children. Development of a rating scale.
. - ,-193 199.

 

94

Luborsky, L. (1962). Clinicians’ judments of mental health. A proposed scale.
AmhixmoﬂgeneraLpsychrm, 407-417.

Masln, E. J., & Johnston, C. (1983). Parental perceptions of child belnavior
problems, parenting self-esteem, and mothers' self-reported stress in younger and older

hyperactive children. IonmaLoLConsultingandLlinimlPsycholoM, 86-99.

Mayer J. E. & Filstead, W. (1979). The Adolescent Alcohol Involvement Scale
(AAIS): An instrument for measuring adolescents' use and misuse of alcohol. Journal

OLSnrdies_on_Alcohol,40,291-300.

Mezzich, J..,E Evanszuck, K..,J Mathias, R..,J Coffman. G.A. (1984).

Admissions decisions and multiaxial diagnosis. AmhmesoﬂGenemljsychratqul,
1001- 1004.

Moretti, M.M., Fine, 8., Haley, G., &Marriage, K. (1985). Child and
adolescent depression: Child-report versus parent-report information IournaLoLthe
AmericanAcarkmy.of.Child.Psychiatry,.24, 298-302.

Mueller, C. & Parcel, T. (1981). Measures of Socioeconomic Status:
Alternatives and Recommendations. ChﬂdDeyelopmensz, 13-30.

Newman, F. & Rinkus, A. (1978). Level of functioning, clinical judgment,

and mental health service evaluation. EvaluatroaaantheHealthﬂnfessronua),
175-194.

Newman, F, (1985). Level of functioning scales: Their use in clinical
practice. Innorzationsjnﬁlinicallracticemrcejook.

Nurcombe, B. (1989). Goal-Directed Treatment Planning and Prinicples of

Brief Hopsﬁliaation.1oumaLof_AmericarLAcademy_of_ChihLamLAdolescent
Psychiatry,_28,_26-30.

Patrick, C. (1993). Use of inpatient services by a national population: Do

beneﬁts make a difference? Ioumal_o£the_American_Academy_of_Child_and_Adolescetn
Psychiatrylﬂ, 144-152.

Puich-Antich & Ryan (1995). TheScheduleionAffecﬁxeﬂsordersand
. , ~ . _- - . . - Unpublished

 

Quay, H. C. (1977). Measuring Dimensions of deviant behavior: The behavior
problem checklist. IonmaLoLAbnormal£hild_Psychology,j, 277-289.

95

Rey,J..,M Plapp, J..,M Stewart, G, Richards,I., &Bashir, M. (1987).
Reliability of the psychosocial axes of DSM-III 1n and adolescent popluation. British

IoumalnfksychnatrgnJiﬂ, 228-234.

Rey,J..,M Stewart, G.W., Plapp, J.M., Bashir, M..,R &Richards,l..N
(1988). Validity of Axis V of DSM-III and other measures of adaptive functioning.

ActaPscychiatnaSmndanidLJI, 535-542.

Reynolds, W. M. & nGraves, A. (1989). Reliability of children’s reports of
depressive syptomatology. _Ioumal_of_AhnonnaLChild_Psychology,_LL 647-655.

Ross, D.M. & Ross, SA. (1982). Hyperactivity. NY: Wiley.

Rotharn, D., Sorrells, J., Heldarnan, P. (1976). A_Global_Assessment_Smle_for
Chldren. Oakland, California, Alameda County Mental Health Services.

Sacco, W. & Graves, D. (1985). Correspondence betweern teacher ratings of

childhood depression and child self-ratings.IoumaLof_Clinical_ChilrLEsychOlogy,_14,
353-355.

Sattler, J .M. (1988). AssessmentoLChildren. Jerome Sattler Publications, San
Diego, CA.

Shaffer, D. Gould, M..,S Brasic, J., Ambronsini, P., Fisher, P., Bird, H. &
A1uwahlia,S. (1983). A Children' 3 Global Assessment Scale (CGAS). Archixeaof

GeneralPsychiatryAﬂ, 1228-1231.

Shaffer, D., Cambell, M., Cantwell, D., Brandley, S., Carlson, 6., Cohen,
D., Denckla,.,M Frances,.,A Garﬁnkel,.,B Klein,.,R Pincus,.,H Sptizer,.,R
Volkrnar, P., & Widiger, T. (1989). Child and adolescent psychaitric disorders 1n

DSM-IV: Issues facing the work group. ChrldandAdolescenLBsychiatqr,_53,_830—
835.

Sorensen, J. L. Hargreaves, W. A. & Friedlander, S. (1982). Child Global
Rating Scales: Selecting a measure of client functioning in a large mental health

system. Evaluationandegramﬂanninm, 337-347.

Sorrels, 1., Rothman, D., & Heldman, P. (1976). A global assessment scale
for children. Unpbulished manuscript. Aladeda County Child and Family Mental
Health Services. Oakland, CA.

Spitzer, R. L., Endicott, J., & Robins, E. (1975). Clinical criteria for
psychiatric diangosis and DSM-III. AmericanJoumaLoﬂPsychiatry,_132, 1187-1192.

96

Sternhausen, H. (1987). Global Assessment of Child Psychopathology.1ournal
_ ‘ , . .- . -, 203-206.

 

Stranger, C. & Lewis, M. (1993). Agreement among parents, teachers, and
children on internalizing and externalizing behavior problems. IoumaLoLCliniral

ChildPsychologLJZ, 107-115.

Taube, C.A., & Barrett, S.A. (1985). MentaLHealth, United States.
Rockville, MD: National Institute of Mental Health.

Vandvik, I. H. (1990). Mental Health and psychosocial fnctioning 1n chilren

with recentonset of rheumatic disease. Ioumal_of_Child_Psychologyand_Psychiatgr,31,
961-971.

Verhulst, F. C. & Van der Ende, J. (1992). Agreement between parents'

reports and adolescents self-reports of probem behavior. Ioumal_of_Child_Psychology
andPsychiatry,_33_(6),1011- 1023.

Weinstein, S. R., Noam, G. G., Grimes, K. ,Sto,ne K. & Schwab-Stone, M.
(1990). Convergence of DSM-III diagnoses and self-reported symptoms in child and

adolescent inpatients. American.Academy.of.ChildaanAdolescenLPscyhi-‘ntq,29,
627-634.

Weissman, M. M. ,Warner, V. & Fendrich, M. (1990). Applying impairment

criteria to children's psychiatric diagnosis. IoumaLoLtheAmericanAcademprhild
aranAdolescenLPsychiatgaﬂ, 789-795.

Weissman, M. Sholomskas, D. & John, K. (1983). The assessment of social
adjustment: anupdate. MhmesoLGeneralPsychratqr,_38,_1250-l258.

Zucker, R., Noll R., Fitzgerald, H. (unpublished instrument). TheMichigan
I: II 0 . .

 

HICHIGQN STRTE UN

II I II uh

3129301579

LIB
”MW
97