. Eva?

. J . ‘ .

.-v.xo|...._ul..h..u! .

. . fifln...‘ m -. ..
:whuflvflul v. I vl 11 D.

h dl!‘§< ‘0‘
g. . I

Wail #- thnlfihwwwrzdh. I

01 t .. {:hYI‘OI .

In

.

v... a o.

.1“. |r§551t€¢233§11
‘t.’ ‘ uni tD. l7].
. T. .wv-L.‘....l§..Z.¢.vu\?
. h :‘l.".- .HII‘HJI,

a! t.» nv‘. .

q .in
v u' 1"

9 I 5.]. .uvll
... . a -

. - .~ .
1.:
u . a I II I «v I Q
I C I o 1:0 I~ A. 1‘ . . ul .0
t a . I 1 .0. A y I u
U v a u. I v u '4': I 1: 0w L.\o.\ o ‘ w . N vac I 7..-.
l I l in! u'1 I r A v 0.11.“... VI(-‘ V1 I '- o‘llv 5‘1 ua¢I\l-I . 0". .
‘ I. c .u .M . u v .y s .III v. to. . .v1.Jn§.v .Hol lat. .v.
V! A t. J~IAI .' 1. ..l . l' 0' ‘ .II I -‘btilOo | ' . .
a o o. 3 .6 . It ‘ y... L. »’ . u .. o» nu. u o I 7'
n .. h. ~ I... f . u... tun .WD 3 .. u \ . . . t.
\ a I \4IAN..VI - b D. c Own. d .I.' Cd‘ .oO' I
o 1 . v-v 1 a v 0.1..
.a . J. .0 V.|,l . , .
1 v \V-.. n IL-..
1.
1..

.
3‘1... .
..I‘ ‘In AO‘L‘ ll
3! . . ..u..Wl..oJr.........
. .\ Mo.tvi0.t "an“
I } "(ll .‘ u n I.
I - . .. f)
3.11:4. v..1lt.oh1..l|"fouhwlco 7“an ”A" .
gimm!‘ Iv..b.ln-nfi¢q.0l||u|0\. vs .0 .
.a tau! 11v. $§I.I.Iltn.lt1li v .v. a .V
:91- . fixhfiu Ezawuf: . . t. - ‘ l -. .
c - urn. (ll 1. vbl‘ “.‘olllf'o I'D}. o on $1 v-v .‘n
. 1 03‘1‘ v ‘5 . » lug-I... I n . . V.-
..Mz g f- whamxurzunuy.-..ul ...... 4.- z: I. .I PI. .
ll N o . u 0‘ . .
I. .41 .I. 1“! O? .11 a .0. '“c. I ‘bl. 55.1‘ '
. ca. qtixl‘l-‘l.uhnufl.h.rir.ls $.38 -. if ll}... .04
. u... ..n .01.}. fibgy. . . .l :Vnummmil
.... -. lu..........:fi. virrJar. b... - “NWT...- \
'tn' Hat‘s... I.VV .‘I‘fi.u .c IbI-l 1‘ 0%....
b 1.0:..."- “1%! ‘uuwu‘luhi: Yb- w“. l
. v'un ‘06..qu . 19!...»- A‘.ol\'z\l. nl'I‘roA I «II-v... l I .
- .... aunt.)...h.u...li¢.lhl.ht..vt¥xt1s!i .inl . ...
n u .l- ‘ t‘ubth‘OnCuY‘fllt.‘0II¢.il0§-Su a
.l I. ‘II ll.‘. Ian-div...“ 'vifld ‘Q'ID
U U. u I '“‘\.R v a x I)... t‘l‘ ‘-
0. O'vlli“-..'~| IV?! ‘1' I ‘.
. 111...? _ fivtnflufl'um01‘15H‘I‘1
.Ns ... l. ... 1b.?) .111: a -.\.}|l|to.t
I 9 Q sol b .IOQ‘A lv‘llw. lw.-‘..nl
.. . .1; - “8% nan.-.
“ - ’ o
. .-

§
.\

‘
)
It . SIN:

‘o‘llllnallvllll‘ I. o. Q

It. 1. . a .LllbbL Onnlpnvulmt. I I ‘n‘. I‘lvn

I.'“..I).- .3.“ 0 AA u .u- . '' .Inv s CPIllvlll‘ olc .o
‘u. tuft:

.‘O-

I
tn 0 2o
"thifl~ 1!" ,
A V I000.’l"10\.h.h0.d.7|..1|t9vl‘ll‘t1l"‘ ‘
.. . “lag. I. . IIh l.‘ ‘. .d‘lll....l.\|lb\luu_.
I a . 1. I! «I a 1.0mm
. . .110.‘ OI‘IO“!
.lil. “‘0 U‘lll.‘o‘.ll-.1.
q.
I. I c nl| . 0 I t
r p u 0 I. f. .1;
n . .2. - ..c I- .Mw... - a
I II 0' *9 '16! 4 .00 I
. . - . a. 1... . xvi. Ln. K?l.uduktflfiicltrumn ...
fill! . : '{Il
. “J. 1....H......namm.\.mhflrtnfc.. enunmm
Inloh‘tulW - "I‘bIbI-Icflhtmfllbflflnu.fl. in .
. itltuuflmq law-Ii ummunr ..a1|\u.-n.:la.vrvll dull.
0.. J '. o.l|v.l.|u.m.. .‘ I“.
II c 1‘. l
1: a

.I i
.P.

I H. v
\

.10.th O flout act . L

”Lona?!” I (hu%‘
. u .. .nnaarfinb}. firnu. ..
.. .l .k 4-. I. {.3. fluidlk.“
. v < . n y'.
. I. . Ike-V’s. I .-

. ...«. N. - F ’L)» .Fr.

 

 

l1||ll||lllll||llllllllllllllllllllllllllllzllllllllsllsllllll

 

 

 

 

3 1293 104762
THESIS J ‘"\
mam” '
M24: “2?“: 2: mate
3 University

_______ ._._,. ,_ . ”W“

This is to certify that the
dissertation entitled
Self Reported Symptomatology in

Major Depressive Illness

presented by

Gregory Alan Holmes

has been accepted towards fulfillment
ofthe requirements for

Ph.D. degree in_C_Q_u_n_5_e]_]'_n_g_,_Ed. Psych.

& Special Education

 

[lg/L flCd/w’t 2/26?” 11404

M a ior prole ssor

 

MSU t5 un Affirmamv Action' Equal Opportunity Institution 0.12771

 

 

MSU

 

 

 

RETURNING MATERIALS:
Place in book drop to

 

LIBRARIES remove this checkout from
_—c_- your record. FINES will
be charged if book is
returned after the date
stamped below.
‘ ”fifi“fi9‘85&23

.m ‘3‘ "'7; a:. I :
‘fiifir j (/7 ~‘b-Q

5‘ -_
_._ i
i.” "'1' 1".

. I H ’9 '7' i
‘i-fi’t .1

.
" ._.~ i
r ' ~-
.. .

”Max.-
.L. ,

. . . g'1 5

‘7. ' 2". . R U 4-
' '0 ‘
. i .‘

 

‘.
‘1 i’:
A i
o
q \

 

 

/4//- 37/3

SELF REPORTED SYMPTOMATOLOGY
IN MAJOR DEPRESSIVE ILLNESS

By

Gregory Alan Holmes

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Counseling, Educational Psychology
and Special Education

1983

0..

ABSTRACT

SELF REPORTED SYMPTOMATOLOGY
IN MAJOR DEPRESSIVE ILLNESS

by

Gregory Alan Holmes

The purpose of the study was to determine whether or not a self
rated instrument could differentiate between endogenous and
non-endogenous depression. A 163 item test named the Differential
Diagnostic Depression Scale (DDDS) was developed to measure 29 symptom
categories associated with depression. LNA-

The DDDS was administered to a sample of 100 patients with a chief
complaint of depression at a consortium of inpatient and outpatient
psychiatric centers. The patients were diagnosed as endogenous or
non-endogenous depression based on a combination of values from the
Dexamethasone Suppression Test (DST) and Research Diagnostic Criteria
(RDC). logy

Two approaches were used to construct scales from the 0005 items.
In the a priori approach items were grouped lingo 25 scales based on
content similarity. The internal consistency oiithe scales ranged from
.57 to .88. A principal component solution was used in the second
method of scale constructiontdi; Fifteen factors were extracted with

T‘

internal consistency values that ranged from .32 to .92.

A. discriminant analysis was used to construct a function that
would best separate endogenous and non-endogenous depression for each
approachi‘ Thirtynfive~subject5wfrom“eachwgroup“were*randomlymselected
for~themanalysisiphaje.fivThevdlicrjfllqant :unction constructed by the a
priori approachxécodrectTyl classified 87% 56f theqisgndomlyuflselected
geeups. The discriminant function constructed by the empirical
approach correctly classified 79% of the randomly selected groups.
Cross validation was attempted by using the discriminant functions to
classify the 30 remaining individuals in the sample. The classifica-
tion accuracy was 57% with the a priori discriminant function, and 43%
with the empirical discriminant function.

A second discriminant analysis was conducted, using a select
number of scales from both the a priori and empirical approach. The
second discriminant function based on the a priori approach correctly
classified 73% of the original seventy subjects, and 60% of the cross
validation groups. The second discriminant function based on 'the
empirical approach correctly classified 73% of the original seventy

subjects and 57% of the cross validation groups.

DEDICATION

To Peggy Nay Anderson
and
in memory of
Nash Anderson

11

ACKNOWLEDGEMENTS

In order to successfully complete a project of this magnitude, the
help and support of many individuals and organizations was critical. I
would like to thank the following individuals:

Dr. Hilliam Farquhar, for his inspiration and guidance throughout
the project; Dr. Robert Bielski, for his direction and supervision in
the area of affective disorders; Dr. Gerald Osborn, for his perceptive
comments on the research project and manuscript; Dr. Andrew Porter, for
his statistical consultations throughout the project; Ms. Susan Irwin
for her help with data coding, preparation, and statistical analyses.

Several other individuals played key roles in the completion of
this research. Dr. Christine Shafer was the chief psychiatric resident
during the project and made outstanding contributions in many formal
and informal areas. Chip Nest, Jay Terbush, and Shelley Smithson
functioned as the chief interviewer/raters for the project and it was a
delight to work with them.

Karen Brainard, Dorothy Stengel, and Randy Scheidt were instru-
mental in the day-to-day operation of the study.

I would also like to thank Dr. Ashok Kaul and his staff at Ingham
Medical, particularly Betsy McPhereson, for their tremendous assistance
during the project. Millicent Lane's concise columns in The State
Journal were extremely helpful in disseminating information about the

study to the Lansing area population. Not to be forgotten are the 100

iii

patients who volunteered their time in a very personal way to share
aspects of their depression with our research team.
Finally, deep appreciation to my fellow members of the Dream

Institute for their support during this project.

iv

CHAPTER 1
THE PROBLEM

The classification of major depressive disorders has been one of
the most controversial and important areas of psychiatric research over
the past six decades. Beginning with the early clinical observations
of Freud and Kraepelin, researchers have sought to establish a
classification system of depression that would aid clinicians in their
understanding of the behavior, etiology, and treatment of the depressed
patient. The impetus behind these research efforts is the urgent
nature of the problem, which can be drawn from ‘two observations.
First, the risk of a given individual developing a major depressive
disorder requiring professional intervention is alarmingly high.
Recent research suggests that approximately 23% of all females and 10%
of all males suffer from major depressive illness requiring
professional intervention at least once during their lifetime (Keller,
et. al., 1982). Secondly, if not properly' diagnosed and treated,
depression is often a debilitating and at times life threatening
illness.

Several major advances in the neurosciences in recent years have
significantly affected the diagnosis and treatment of depression-and
serve as a foundation for further research. A recent review of
research attempts to establish valid and reliable criteria for a
diagnostic. system suggests that depressive disorders can be dicho-

tomized into two groups based on the presence or absence of a

l

constellation of behavioral symptoms, biological markers, and treatment
response (Nelson and Charney, 1981). The first group, generally
referred to as endogenous depression or major depression with
melancholia, is relatively autonomous to environmental change.
Furthermore, it is more likely to have neurophysiological correlates,
and is more responsive to biological intervention than non-endogenous
depression. The second group, non-endogenous depression, is typically
defined by the absence of the features mentioned for the endogenous
group.

Two recent developments are of particular importance in attempts
to distinguish endogenous from non-endogenous depression. First has
been the discovery that in 50% of endogenous depressives the admini-
stration of dexamethasone failed to produce normal serum cortisol
suppression. From this finding, the dexamethasone suppression test
(DST) was developed. The dexamethasone suppression test has proven to
be helpful in corroborating the diagnosis of endogenous depression when
it is abnormal (positive), having a specificity of 95% (Carroll, et.
al., 1981). The second development was the introduction of Research
Diagnostic Criteria (RDC) for twenty-five major psychiatric disorders,
including endogenous depression. Spitzer and his colleagues
demonstrated that interviewers using the RDC were able to reliably
distinguish endogenous from non-endogenous depression, with
coefficients for agreement above .80 for both categories (Spitzer, et.

al ., 1978).

Need for the Study

 

A need exists for an additional method of differentiating
endogenous and non-endogenous depression. The majority of research
attempts to classify depressive disorders as either endogenous or
non-endogenous used observer ratings of symptoms, rather than patient
self ratings. Although self rating scales for depression exist,
previous studies have shown that these scales are useful as indexes of
severity of illness and do not differentiate between subtypes of
depression (Rehm, 1976). A self rated scale that could reliably
differentiate subtypes of depression would be of significant value both
as a research instrument, and most importantly, as a relatively
inexpensive technique for the diagnosis of depression.

Both the DST and RDC have limitations which interfere with their
utility as diagnostic procedures for the routine clinical evaluation of
depression. Although the DST has been found to be a relatively
specific laboratory test, 33-50% of endogenous depressives have normal
cortisol suppression after dexamethasone administration and are not
detected by the test (Carroll, et. al., 1981). Additional drawbacks
include the cost of the test and the necessity of having trained
personnel to administer and analyze test results. The RDC are designed
to be used in conjunction with an in-depth clinical interview which is

both expensive and time consuming.

Purpose of the Study

The purpose of the research reported in the present study was to
determine whether a patient self reported instrument could differen-

tiate endogenous and non-endogenous depression.

Research Hypotheses
Gerald Klerman, in his review on research in depression, argues
that the concept of endogenous depression embodies four component
propositions, or hypotheses, that are testable through empirical
investigation (Klerman, 1972). Klerman's four propositions are as
follows:

1. Endogenous depression consists of a pattern of behavioral
signs or symptoms that occur together at one point in
time.

2. The short term episode of endogenous depression, as
contrasted to non-endogenous depression, occurs without a
recent precipitating event.

3. The premorbid personality pattern of the endogenous
depression will be relatively stable and non-neurotic.

4. Endogenous depression will run an autonomous course, i.e.,
will not be reactive to environmental alteration.

In this study an empirical investigation of the first proposition
was conducted. It was hypothesized that:
1. A. self report measure designed t0~ assess empirically
supported constructs will differentiate between endogenous
and non-endogenous depression. -
2. A self report measure will yield interpretable constructs
which will contribute to the understanding of the

differences between endogenous and non-endogenous depres-
sion.

Theory

Although there has been a wealth of research over the past seventy
years on the phenomena of depressive illness, there remains no clear
consensus among investigators regarding the exact classification of
such disorders (Kendell, 1976; Eysenck, 1970). Several conceptual and
methodological problems interfered with attempts at establishing an

accurate nosological system. A major problem was that symptoms

associated with depression, such as disturbance in mood, anxiety, and
decreased concentration occur in a range of psychiatric conditions and
are also part of the normal adaptation response (Klennan, 1972).

One of the fundamental conceptual issues in depression research is
whether depression can be best explained by a unitary or binary model
of classification. Proponents of the unitary model of depression argue
that all depressive illness has basic underlying similarities, and that
all observed clinical differences can be explained by severity of the
illness (Klerman, 1972). In the unitary model distinct demarcations of
subtypes of depression are seen as impossible.

In contrast, the binary, or pluralistic model of depression holds
that it is possible to distinguish between at least two subtypes of
depression, and that differences cannot be explained by severity alone
(Rosenthal and Klerman, 1966). The two most frequent binary models
referred to in the literature are the endogenous-reactive dichotomy and
the psychotic-neurotic dichotomy. hi the former, depressions can be
subdivided based upon their reactivity to environmental change
(Gillespie, 1929). The endogenous depression, once established, is
thought to run an independent or "autonomous“ course, i.e., be
relatively unaffected by external stimuli. As such, these depressions
are often thought to be of biological origin. 0n the other hand,
reactive depressions are so named because of the relative responsive-
ness to external stimuli and are thought to originate as reactions to
life stress.

In the psychotic-neurotic binary model, emphasis is placed on the
presence or absence of psychotic symptoms in making the distinction

between depressions (Kendell, 1976). Psychotic depressions are

characterized by symptoms such as hallucinations, delusions, and degree
of ego regression, whereas neurotic depressions are defined by the
absence of these psychotic symptoms. Unfortunately, as MacFadden
(1963) noted, the labels and descriptive terms for both dichotomies
have been used interchangeably by some researchers, contributing to
confusing and conflicting research reports. For example, some
researchers equated endogenous with psychotic depression, when in fact
the phenomena can occur independent of one (another (Andreasen (and
Grove, 1982). Similarly, the terms neurotic and reactive depression
have been used interchangeably while research suggests the terms have
different implicit and explicit meanings (Akiskal, et. al., 1978).
There are two research approaches that have been used to establish
the validity of the binary, or pluralistic model of depression. One
approach is by demonstrating that subtypes of depression exist, based
upon differences in clinical features, or symptomatology. The second
approach is to demonstrate that identified subtypes respond
differentially to various treatment modalities, such as medication or
psychotherapy. In this study, the former approach is used to study the

differences in the self report of symptoms, as measured by the 0005.

Overview of Remaining Chapters

In Chapter Two, the relevant literature is reviewed in the
following areas: symptom differences, differential treatment response,
and the efficacy of previous patient rated instruments. The research
design and procedures are presented in Chapter Three. In Chapter Four,
the analyses of the results are presented. Conclusions and recommen-

dations for further research are presented in Chapter Five.

CHAPTER 2
REVIEW OF LITERATURE

In the following review, three areas of research are examined.
First, a review of attempts at classifying depression based on symptom
frequency is presented. Second, there is a brief summary of research
on treatment response among subtypes of depression. In the final part
of the review, there is a consideration of attempts of using patient

rated instruments to measure subtypes of depression.

Symptom Frequengnytudies

 

One of the earliest attempts at distinguishing between subtypes of
depression was made by Freud (1956) in his monograph "Mourning and
Melancholia." Freud felt that mourning was a normal response of an
individual to the loss of a loved one, whereas melancholia was a
pathological, or abnormal process. He believed that melancholia could
be distinguished from the mourning process in that melancholia was
associated with strong self reproach, whereas there was no loss of self
esteem in mourning.

Another early attempt to classify depression was conducted by
Kraepelin (1921) who also felt it was possible to distinguish a normal
reaction of sadness from clinical depression. Based upon his clinical
observations, Kraepelin subdivided clinical depression into two types -
psychogenic depression, and manic depressive insanity. The major

characteristics of the manic depressive group were that the attacks of

illness appeared to have no precipitant, and that the course of illness
was “independent," or autonomous, from events that occurred during
their hOSpitalization. Psychogenic depressions were believed to be
precipitated by external circumstances and influenced by the milieu
during the course of treatment.

Gillespie (1929) reported his clinical observations of a group of
25 depresed inpatients who were given physical and mental examinations.
He found that their depressions could be differentiated according to
degree of reactivity to external and internal stimuli. Two major
groups were discovered: reactive and autonomous depressions, with the
latter having a subgroup which was labeled as "involutional." The
reactive depressions were primarily characterized by presence of a
precipitant and their responsiveness during the course of treatment in
the hospital. Approximately 90% had clearly identifiable events which
were thought to precipitate the patients presenting problems.
Furthermore, the reactive depression group was judged to be responsive
to their physician or other variations in the hospital milieu, such as
visits by family members, or staff changes. Gillespie observed that
this group was also more responsive to everyday events such as changes
in the weather.

The autonomous group was characterized by a lack of responsiveness
to most of the events mentioned above. However, Gillespie was able to
identify a precipitant for the majority of the patients in the
autonomous group, and concluded that the distinctive feature of this
group was the autonomous nature of the illness once the depression had
been activated. Other characteristics of the autonomous group were

increased “restless activity,“ feelings of remorse, frequent self

accusations, and a family history of psychoses. A small subgroup of
the autonomous depressions were labeled as involutional, and exhibited
a preponderance of hypochrondriacal concerns and poor insight.

Lewis (1938) challenged the early attempts at subdividing
depression and claimed that depression could be classified according to
severity alone. He stated that the manic depressive illnesses most
likely represented an acute, severe depression, whereas the other
clinical depressions were “...chronic, mild depressions." Most
importantly, he argued that differences in symptoms could not be used
in a meaningful way to separate subgroups.

There have been several attempts to discriminate endogenous from
non-endogenous depressions using factor analysis of symptomatology.
Hamilton and White (1959) used factor analysis on 17 clinical features
in a group of 64 depressed male inpatients. Four factors were
extracted which they interpreted as "retarded depression," “agitated
depression," "anxiety reaction," and "psychopathic depression." The
first factor was felt to represent an endogenous depression, and the
symptoms with the highest loading on the factor were depressed mood,
self reproach, psychomotor retardation, suicidal ideation, and loss of
interest in activities.

Heckowicz, Cropley, and Muir (1971) attempted to replicate
Hamilton's findings in a study of 52 depressed males on an inpatient
unit; in Canada. Four factors were found in the factor analysis.
Comparison of the results with the Hamilton results indicated little
similarity between the two sets of factors, and none of the factors
resembled the ”retarded depression” found in the earlier study. The

authors suggested that the relatively small sample sizes in both

10

studies could be a major reason for the lack of agreement between the
studies.

Kiloh and Garside (1963) conducted a factor analysis on 35
clinical features noted as present or absent in 143 cases of depressed
outpatients. Two factors were extracted, the first of which was
interpreted as a general factor; the second, a bipolar factor. The
second factor was found to have a high correlation with the clinical
diagnosis of endogenous or neurotic depression. The clinical features
with the highest correlation with the diagnosis of endogenous depres-
sion were early awakening, morning worsening, distinct quality of mood,
and psychomotor retardation. Neurotic depression was characterized by
reactivity to the environment, presence of a precipitant, self pity,
and variability of depressed mood. They argue that their results
support the “...traditional dichotomy of depressive cases into neurotic
and endogenous varieties..."

An attempt to replicate Kiloh and Garside's original study was
conducted by Kiloh and his associates (1972) using 145 inpatients in
Australia. Principal component analyses on the data fron the original
study and the replication produced two factors which were similar in
nature. The first factor was a bipolar factor which the authors
believed to be descriptive of endogenous depression. Unlike the first
study, no general factor was discovered, and was thus considered to be
an artifact of the analysis technique used in the first study. The
second factor was thought to be descriptive of neurotic depression.
Features associated with the endogenous factor were depth of depres-

sion, early awakening, psychomotor retardation, and quality of the

11

depression. The neurotic factor was characterized by reactivity,
presence of a precipitant, duration, and inadequate personality.

The authors speculated that, whereas endogenous depression is
likely to have the quality of a categorical illness due to the tendency
for the scores to cluster together, neurotic depression was more likely
to be a dimensional illness. Individuals with neurotic depression
would not be expected to exhibit endogenous features, whereas
individuals with endogenous depression may respond to their illness
with neurotic behaviors.

McConaghy, Joffe, and Murphy (1967) attempted to replicate Kiloh
and Garside's 1963 study. One hundred outpatients were evaluated for
the presence or absence of 43 clinical features, 35 of which were
included in the Kiloh and Garside study. Two factors were extracted,
and the authors concluded that the factor loadings for each clinical
feature did not indicate that either factor could differentiate
neurotic from endogenous depression. Eight features, thought to be
most representative of endogenous and neurotic depression were selected
for factor analysis, and neither of the two factors that were extracted
were associated with either type of depression.

The authors attributed the differences between their results and
the Kiloh and Garside study to possible interviewer bias in Kiloh's
study and different methods of patient selection. The interviewers in
the Kiloh study were aware of the clinical diagnosis during the
interview. Secondly, Kiloh and Garside performed their analysis on the
third of their sample that they believed had “definite" diagnoses,

whereas, the HcConaghy study did not exclude patients from the

analysis.

12

Rosenthal and Gudeman (1967) conducted a factor analysis of the
clinical features of 100 female outpatients and inpatients and
extracted a bipolar factor which they interpreted as endogenous.
Symptoms with the highest loading on the factor were lack of
reactivity, concentration difficulties, distinct quality of mood,
psychomotor retardation or agitation, and midnight awakening. Symptoms
with a negative loading on the factor were self pity and irritability.

Kay, Garside, Beamish, and Roy (1969) confirmed the existence of a
bipolar factor in their study of 104 of inpatients selected retro-
spectively fron hospital records for a five year follow-up study.
Thirty-five features were rated as absent or present, based on the
hospital records. Symptoms with negative signs on the factor were
guilt, psychomotor retardation, severity of depression, hopelessness,
suicidal behavior and nihilistic ideas. The authors concluded that
this cluster was similar to the endogenous syndrome described by
Hamilton and White (1959), Kiloh and Garside (1963), and Rosenthal and
Gudeman (1967). However, it should be noted that the 104 patients
represented 29% of the cases eligible for analysis, with 71% rejected
due to diagnosis, incomplete data, or inability to contact the subject.

Garside, Kay, Hilson, Deaton, and Roth (1969) rated 269 depressed
inpatients on the Depressive Category Type Scale, (DCTS), which was
composed of 15 items believed to distinguish endogenous from neurotic
depression. A principal component analysis was conducted and the
distribution of the first component scores was calculated. Both the
distribution of component scores and the sum of unweighted raw scores
had a bimodal distribution, leading the authors to argue that the

patients could be divided into an endogenous group and a second group

13

characterized by the absence of endogenous features. The features with
the greatest weight on the first component were autonomous course,
psychomotor retardation, distinct quality, good premorbid personality,
and morning worsening.

Klein (1974) observed that most factor analytic studies have shown
a bipolar factor that cannot be reduced to a severity factor alone. He
suggests that research, in attempting to demonstrate a binary view of
depression, has been handicapped by researchers using different
definitions of endogenous depression. Klein notes some researchers
believe endogenous depression refers primarily to a lack of a
precipitant, whereas others use endogenous depression to refer to the
lack of reactivity of the depression to environmental changes. He
observes, as have others, that many so called endogenous depressions
have been found to have a precipitant whereas others have not. He
suggests using the term "endogenonorphic depression” to describe the
symptom pattern traditionally associated with endogenous depression, as
the term endogenomorphic does not necessarily imply a biological
etiology. Furthermore, he believes that endogenomorphic depressions
can then be subdivided into two types: endogenous depressions, and
precipitated depressions with endogenous features. In making the
distinction, Klein implies that the endogenous-reactive conceptuali-
zation of being polar opposites is incorrect, and that the two concepts
are independent from one another.

Lewinsohn, Zeiss, Zeiss, and Haller (1977) conducted a factor
analytic study that supported Klein's hypothesis that endogeneity and
reactivity are independent dimensions. Three groups of depressed

outpatients were assessed on 35 symptoms, and a separate factor for

14

endogeneity and reactivity was identified in all three samples. No
bipolar factor was extracted, lending support to Klein's two dimen-
sional model. Symptoms with the highest loading on the endogeneity
factor were helplessness, self reproach, psychomotor retardation, lack
of reactivity, loss of interest, and distinct quality of mood.

Daly and Cochran (1970) investigated the occurrence of clinical
features in 241 white females admitted to the inpatient unit of a North
Carolina Hospital. Individuals with a diagnosis of involutional
melancholia were compared with all other diagnoses, based on the
presence or absence of 31 clinical features. Eight of the 31
comparisons proved significant by chi square. Those individuals in the
involutional group were more religious, compulsive, had fewer
difficulties in their adult lives, were more likely to receive ECT,
antidepressants, and supportive psychotherapy, and exhibited either
psychomotor agitation or retardation. A second analysis compared
subjects classified as endogenous depressives with those diagnosed as
reactive depressives, and 18 comparisons were reported as significant.
The endogenous group was described as more religious, less likely to
abuse alcohol, and have fewer passive aggressive features or childhood
and adult difficulties.

Hatussek, Soldner, and Nagel (1981) attempted to confirm the
existence of an endogenous syndrome by analyzing symptom frequency and
using cluster analysis on the symptoms of 198 subjects who had been
previously hospitalized for depression. The subjects were diagnosed as
either endogenous or neurotic by Research Diagnostic Criteria. In
comparing the symptom frequency between the two groups, the endogenous

group was significantly related to eight symptons: morning worsening,

15

lack of reactivity, short duration, distinct quality of mood, psycho-
motor retardation, indecisiveness, sudden onset, and delusions. The
symptoms associated with neurotic depression were sadness and
neuroticism.

The cluster analysis determined eight symptoms as being characte-
ristic of the endogenous syndrome: distinct quality of mood, lack of
reactivity, withdrawal from social contact, impulse inhibition,
disturbance of the circadian rhythm, physiological disturbance
(appetite or sleep), sudden onset, and absence of a precipitant. They
conclude that it is possible to detect an endogenous syndrome by
cluster analysis, and believe that the syndrome cannot be defined by
the presence or absence of a single item. They suggest the presence of
an "endogenous component" with varying levels of strength in all
depressive illness.

Andreasen and Grove (1982) used cluster analysis to classify a
sample of 275 depressed inpatients. One hundred and six symptoms were
rated based on the results of a semi-structured interview. Information
on course of illness, family history, and treatment variables were used
to validate the clusters. Four groups were found as follows: a
severely depressed group with endogenous features, a moderately
depressed group, a bipolar group, and a depressed group with psychotic
features.

The authors concluded that the results indicate that subtypes of
depression exist and differ in terms of both severity of illness and
profile of symptoms. They stated that depressive disorders are

”...probably not a unitary and homogenous phenomena...", and that the

16

results were independent validation of the subtypes described by
Research Diagnostic Criteria.

Feinberg and Carroll (1982) have recently reported a successful
attempt at separating unipolar endogenous depression from
non-endogenous depression using discriminant analysis. One hundred and
sixty-five inpatients and outpatients were diagnosed as endogenous or
nonendogenous based on material from the Schedule for Affective
Disorders and Schizophrenia, clinical interviews, and response to
treatment. Each patient was rated on several clinical and historical
features, as well as the Hamilton Rating Scale. The ratings for all
patients with an HRS score over 10 were used in a discriminant analysis
in an attempt to validate clinical diagnosis. A discriminant function
that separated the two depressed groups contained eight items:
decreased appetite, guilt, presence of a precipitant, agitation,
delusions, work and interests, retardation, and loss of pleasure.
Eighty-four percent of the patients were classified correctly using the
discriminant function.

Matussek and Luks (1981) found significant differences in the
thematic content of endogenous and non-endogenous depressives. Fifty-
three women with past episodes of depression were classified as having
had endogenous or non-endogenous depression according to Research
Diagnostic Criteria. Interviews were conducted with each subject to
determine the themes, or concerns of the depression, and to determine
whether or not the subject could see a connection between their illness
and the contents mentioned. Three themes were reported more frequently
in the non-endogenous group: concerns over separation from significant

others, arguments <n~ conflicts with other people, and self reproach.

17

Furthermore, endogenous patients more often found their depressive mood
inexplicable, i.e., could not see the connection between their problems
and mood. Matussek and Luks concluded that the endogenous depressive
may be less capable of establishing and maintaining intensive emotional
relationships.

Nelson and Charney (I980) evaluated the frequency of primary
affective disorder in subjects with endogenous depression and reactive
depression in an attempt to use that criteria to reliably distinguish
between the two groups. One hundred and two inpatients were classified
as endogenous or reactive depending upon their responsiveness to
psychosocial aspects of hospitalization. (Hdteria for primary
affective disorder were taken from Research Diagnostic Criteria, and
the subjects were also rated on symptoms for the RDC subtype of
endogenous depression. The composite primary affective disorder
criteria defined a heterogenous group of patients, i.e., included a
significant percentage of both endogenous and reactive depressions.
However, significant differences were found between ther two groups
among individual symptoms, with agitation, retardation, self reproach,
decreased concentration, and depressive delusions occurring more
frequently in the endogenous group. The presence of a precipitant did
not distinguish between the two groups.

Although many of the above investigations appear to support an
endogenous/non-endogenous model of depression, there is less support
for the psychotic-neurotic model. Akiskal and his associates (1978)
conducted a four year prospective study in an effort to determine the
homogeneity of patients that had been given the diagnosis of neurotic

depression. During a four year follow-up period, 40% of the sample

18

were diagnosed as having a primary affective disorder, whereas 48% were
diagnosed as a secondary affective disorder and 12% fit neither
category. Furthermore, fully 70% of those diagnosed as having primary
affective disorder exhibited endogenous features as defined by RDC.

The authors argued that the results support the conclusion that
the term neurotic depression may no longer be useful as it refers to a
heterogenous group of disorders. Secondly, since all of the subjects'
depressions were precipitated by an event, they believed that it was
not feasible to use precipitants as a criterion for subtypes of depres-
sive disorders. Finally, the results supported Klein's concept of
endogenomorphic depression as 36% of the sample developed depressions
with endogenous features.

Klerman's study (1979) of depressed inpatients supports the
finding that the) diagnosis. of neurotic depression lacks sufficient
clarity to be used as a research classification. Klerman identified
six criteria as forming the constellation of neurotic depression: mild
severity, lack of psychotic features, presence of a precipitant,
presence of longstanding maladaptive personality patterns, and the
presence of unconscious conflicts. Ninety depressed patients with
major affective disorder were diagnosed with the RDC to determine the
degree of overlap between four of the above criteria. Only 17% of the
sample met all four criteria, and the degree of overlap between any two
symptoms varied between one-half and two-thirds. A second major
finding from the study was that 37% of the patients with endogenous
features had a clearly defined precipitant. Klerman concluded that the

findings demonstrated that the multiple criteria for neurotic

19

depression defined different groups of patients and therefore were too
vague for clinical and research use.

Kendell and Gourlay (1970) attempted to determine if psychotic
depressives and neurotic depressives could be separated using discri-
minant analysis. One hundred and fifty features were gathered on 63
patients with depressive neurosis and 115 patients with the diagnosis
of psychotic depression. Twenty-four of the 150 items discriminated
between the two groups in a series of chi-square tests, and these items
were then combined with an additional nine items for the discriminant
analysis. Although most of the psychotic group had positive scores and
most of the neurotic group had negative scores on the discriminant
function, the distribution was not significant. Interestingly, Kendell
and Gourlay found the overlap between the two groups to be smaller than
a previous study they conducted (Kendell, 1968) and attribute this
difference to more consistant diagnostic criteria and higher
reliability of their data.

Several previous reviewers of the literature have concluded that
there is considerable research support for a binary model of depres-
sion. In her review of the major statistically based classification
studies, MacFadden (1963) cautioned against drawing generalizations
from interstudy comparisons. She noted that often the reliability of
clinical diagnosis in the studies was low and varied according to the
setting in which it was conducted. Furthermore, she observed that
considerable semantic differences existed in that investigators used
such diagnostic labels as “endogenous" to describe different phenomena.
However, MacFadden concluded that given these reservations, factor

analytic, cluster analytic, and discriminant function studies all point

20

to the existance of a "...persistant, severe, 'endogenous' depression."
0n the other hand, she concludes that the evidence was less supportive
for the existence of a "specific" neurotic or reactive depression,
although several studies suggested the presence of two or more types.

In his review, Roth (1960) concluded that endogenous and neurotic
depression can be differentiated according to specific symptons and
premorbid personality. He believed that in endogenous depression the
affective change has a "...disproportionately severe depth and
intensity” in relation to the precipitant, and that patients report a
distinct quality of mood. Other outstanding features of endogenous
depression were lack of responsiveness to environmental change, early
morning awakening, psychomotor change, and delusions.

Eysenck (1970) emphatically stated in his review of the literature
that the debate between proponents of a unitary model and binary model
"...has been conclusively decided in favor of the binarians." There-
fore, he has suggested that researchers should concern themselves with
whether or not depression can best be explained by a categorical or
dimensional model.

Kendell (1976) was far more reserved in his review. He noted that
semantic differences made any comparison of studies difficult at best.
In an awkward attempt to circumvent such differences, he discussed the
literature in terms of support for two subtypes which he labeled "Type
A" and “Type 8". Type A depressions were considered to be severe
depressions with the symptoms of diurnal variation, guilt, retardation,
insomnia, and weight loss. Type B depressions were described as milder
in severity and without the symptoms associated with Type A

depressions.

21

Although Kendell felt that the controversy between the unitarians
and binarians was not completely resolved, he did concede in the review
that there was a need to separate the types and that most of the
research supports the notion of Type A depression or syndrome. He
believed there was much less agreement on what constituted the Type B
syndrome.

Nelson and Charney (1981) concluded from their recent review of
the literature that autonomous or endogenous depression could be
differentiated from non-endogenous depression by the presence or
absence of specific symptoms. Further, they suggested 'that ‘there
appeared to be two subtypes of endogenous depression, an anhedonic,
retarded type and an agitated, delusional type. The symptoms with the
greatest association with endogenous depression were psychomotor
change, severity of depressed mood, lack of reactivity, depressive
delusions, self reproach, and loss of interest. They believed that
there was moderate research support for the association of the
endogenous symptoms of distinct quality of mood, diurnal morning
worsening, and difficulty concentrating. Sleep disturbance, weight
loss, appetite disturbance, and suicidal thoughts were not found to be
helpful in differentiating between subtypes of depression.
Furthermore, they argued that the symptom differences may prove to be
useful in the generation of valid diagnostic criteria, and propose that

further research was needed to determine their reliability.

Treatment Response Studies
Several researchers have indicated support for a binary model of
depression based on studies of response to treatment. Garney, Roth,

and Garside (1965) examined the relationship between symptoms of

22

depression, clinical diagnosis, and response to electroconvulsive
therapy (ECT). One hundred and twenty-nine inpatients who were
referred for ECT were rated for the presence or absence of 35 clinical
features, and followed for six months after ECT. A significantly
higher percentage of patients diagnosed as endogenously depressed
responded to ECT than those diagnosed as neurotic depressive. Three
significant factors were found through factor analysis of the features:
a bipolar factor, corresponding to endogenous and neurotic depression,
a general factor, and a paranoid psychotic factor. High positive
loadings of the bipolar factor were found for adequate premorbid
personality, absence of a precipitant, distinct quality of mood, weight
loss, body build, history of depression, early' morning awakening,
psychomotor change, somatic and paranoid delusions, and ideas of guilt.
Negative loadings were found for the following symptoms: anxiety,
morning worsening, self pity, and-hysterical features. The authors
considered these results verification of the hypothesis of two distinct
depressed populations.

Raskin and his colleagues (1970) investigated the response of
depressed subgroups to chlorpronazine, imiprimine, and placebo in a
double blind study of 555 depressed inpatients. Three subgroups were
formed based on initial diagnosis at the time of hospitalization:
neurotic. depressives, psychotic depressives, and schizophrenic
depressives. Clinical status was evaluated prior to drug administra-
tion and at selected intervals thereafter. No difference in response
to medication were found among neurotic depressives. Imiprimine was

found to be more effective that either placebo or chlorpromazine in the

23

psychotic group. Raskin concluded that differential response supports
distinguishing among depressive subgroups.

Indirect support for findings of differential treatment comes from
the work of Gurney and his colleagues (1970). One hundred and fifty-
four inpatients with primary affective disorder were assigned to two
groups according to whether anxiety or depression was the predominant
affect during a clinical interview. Response to treatment was measured
at discharge and at a six month follow-up. Electroconvulsive therapy
and tricyclic antidepressants were prescribed more frequently to the
group with depression as the predominant state, whereas more sedatives
and barbituates were prescribed to the anxiety group. In addition, the
depressed group had a significantly better response rates to both ECT
and tricyclic medication. The authors believe that the findings
demonstrated the importance of separating affective disorders according
to whether or not anxiety was the predominant symptom.

Overall and his associates (1966) investigated the relationship
between symptoms and diagnosis during a study on the effects of
medications on depressed subjects. Seventy-seven patients were
classified on 16 symptom variables into three profile clusters through
factor analysis. These three profiles were interpreted by the
investigators as "anxious-tense depression," a "hostile depression,"
and a "retarded depression.“ The anxious-tense depression was thought
to correspond with what has been called reactive, or neurotic
depression, whereas the retarded depression corresponded with features
typically seen in endogenous depression. Differential treatment
response to medication was evaluated between the three groups, and the

subjects with ”anxious-tense" depression responded significantly better

24

to thioridazine, whereas the “retarded depression" group had signifi-
cantly greater improvement to imiprimine. Overall suggests that
depression can and should be subdivided according to the presenting
profile of symptoms for treatment response.

Two reviews of treatment response studies indicate support for the
notion of differential response of symptom groups to different treat-
ment. Nelson and Charney (1981) concluded in their review that
psychomotor retardation, loss of interest and emotional withdrawal
predict response to tricyclic depressants. Bielski and Friedel (1976)
concluded in their review of the literature on prediction of tricyclic
response that anorexia, weight loss, and middle or late insomnia were

most predictive of response to medication.

Self Report Measures

 

Although there appears to be a growing consensus among researchers
that a binary, or pluristic model may best describe depressive

disordersA questions remain as to how to best measure symptomatology.

(L’-
i '1 _li{7v{')‘fl‘~

The majority of research; has used interviewer ratings of symptoms
5v ~94;th CNN “4

rather than the self report of patients. Although several self rated
measures of depression exist, Rehm (1976) found that few of the scales
had psychonetric data to support their use. Rehm concluded that the
Beck Depression Inventory, MMPI, Zung Depression Scale, and Zubin
Adjective Check List had the most psychometric support and are the most
frequently used, however, each measured a limited range of depressive
symptoms.

Hughes, O'Hara, and Rehm (1982) reviewed the most conlnonly used

self rated instruments that are used to measure depression and found

that many of the instruments were limited by the restricted content

“P

l
l
‘.
l

l

__.—_-.
.——_._.._-——-’

25

they sample. They report that the Beck Depression Inventory emphasized
the cognitive disturbance of the subject, whereas the Hamilton Rating
Scale emphasized physiological correlates. The authors believe that
self report questionnaires have several advantages over other measures
of depression, including relative inexpense, ease of administration,
and their ability to quantify subjective complaints. The disadvantage
of self report questionnaires is that they are susceptible to cultural
and subjective response bias. The'a‘u‘tqhors concluded after comparing '
self report questionnaires to other techniques that no single class of i
instrument could be considered best and recomended combinations of
instruments to establish clinical diagnosis. C€:;::;§&i~

One of the most frequently cited studies on the validity of self
report scales was conducted by Prusoff, Klerman, and Paykel (1972).
Two hundred depressed subjects were evaluated by interviewers using the
Hamilton Rating Scale and a 110 item self report inventory. Evalua-
tions were conducted at the time of hospitalization with correlations
ranging from .11 to .63. At follow-up the correlations were somewhat
higher, ranging from .34 to .74. Correlations for severity were .83.
The authors suggested that whereas self report assessments may be
helpful in measuring the presence or absence of symptoms, they cannot
be used to measure severity of illness during the acute episode.

Although the Prusoff study has been cited by some as evidence for
the limited usefulness of self report measures, there are several
methodological flaws in their study, some of which they frankly
acknowledge. One important flaw was that the instruments differed in
content, and the wording for various symptons areas was different.

Secondly, two different raters were used at time of admission and

26

follow-up, and no data is presented on inter-rater reliability.
Finally, only the data from 67% of the sample was reported as fully
one-third of the sample did not participate at follow-up.

Feinberg and his associates (1981) reported a higher concordance
rate when similar instruments were used by subjects and raters.
Comparisons were reported between scores on the Hamilton Rating Scale.
and the Carroll Rating Scale, a self rated version of Hamilton.
Correlations of .75 were found between the total scores on a sample of
198 depressed outpatients. Furthermore, Feinberg's group reported
differences in the correlation coefficients between subtypes of
depressed subjects, with those patients diagnosed as unipolar
depression having a correlation of .83, bipolar depression with a
correlation of .75, and non-endogenous depression of .66. The authors
suggested that patients with non-endogenous depression may report more

symptoms than are rated by clinicians.

Summary and Conclusions

Several conclusions can be drawn from the preceding review of the
literature. As MacFadden (1963), Kendell (1976), and Nelson and
Charney (1981) have noted previously, considerable differences do exist
among the earlier studies in sample composition, definition of key
clinical diagnostic tenns, and symptoms included. However, the clear
majority of the previous studies lend support to the conclusion that
depression can best be explained and treated by using a binary, or
pluralistic paradigm. The twenty-two major studies reviewed in this
chapter are presented in Table 2-1. Only three of these studies failed

to support a binary model of depression.

27

Table 2-1

Summary of Symptoms

Associated with Endogenous Depression

 

 

Author(s)

1. Freud

2. Kraepelin

3. GilleSpie

4. Lewis

5. Hamilton and White

11.

12.

Kiloh and Garside

Garney, et. al.

Overall, et. al.
McConaghy, et. al.

Rosenthal and Gudeman

Garside, et. al.

Kay, et. al.

Year

 

1917
1921

1929

1938
1959

1963

1965

1966
1967
1967

1969

1969

Symptoms

Self reproach

Lack of precipitant, autonomous
course

Lack of reactivity, psychomotor
agitation, self reproach, family
history of psychoses.

No qualitative difference.

Severity of mood, self reproach,
psychomotor agitation, suicidal
ideation, loss of interest.

Early awakening, diurnal varia-
tion, quality of mood,
psychomotor retardation.

Premorbid personality, absence
of precipitant, distinct quality
of mood, weight loss, body type,
family history, early morning
awakening, psychomotor change,
delusions, self reproach.

Responsiveness to imiprimine.
None.

Lack of reactivity, poor concen-
tration, distinct quality of
mood, psychomotor agitation or
retardation, midnight awakening.

Autonomous course, psychomotor
retardation, quality of mood,

premorbid personality, diurnal
variation.
Guilt , psychomotor retardation ,

severity, hopelessness, suicidal
behavior, nihilistic ideation.

13.

14.

15.
16.
17.

18.

19.

20.

21.
22.

Daly and Cochrane

Gurney, et. al.

Raskin, et. al.
Heckowicz, et. al.

Kiloh, et. al.

Lewinsohn, et. al.

Nelson and Charney

Matussek, et. al.

Matussek and Luks

Andreasen and Grove

1970

1970

1970
1971
1972

1977

1980

1981

1981
1982

28

Religious behavior, compulsive,
psychomotor agitation or retar-
dation.

Responsiveness to ECT and
tricyclic medication.

Responsiveness to imiprimine.
None.

Severity of depression, early
awakening, psychomotor retarda-
tion, quality of mood.

Helplessness, self reproach,
psychomotor retardation, lack of
reactivity, loss of interest,
distinct quality of mood.

Psychomotor agitation or
retardation, self reproach,
decreased concentration,
delusions.

Distinct quality of mood, lack
of reactivity, social with-
drawal, impulse inhibition,
sleep and appetite disturbance,
sudden onset, lack of precipi-
tant.

Loss of insight.
Insomnia, loss of appetite, loss

of interest, psychomotor agita-
tion, lack of reactivity.

 

29

The frequency of the association of endogenous depression with
specific symptoms is summarized in Table 2-2. Psychomotor disturbance
was the most frequently reported symptom, appearing in 12 of the 22
studies reviewed. Other frequently reported symptons were automous
course, self reproach, distinct quality of mood, and middle or late
insomnia. Symptoms that were reported relatively infrequently as
distinguishing endogenous from non-endogenous depression included
sudden onset, withdrawal, suicidal ideation or behavior, family
history, appetite disturbance and diurnal variation. Finally,
responsiveness to tricyclic medication and ECT appears to be more
characteristic of endogenous depression than non-endogenous depression.

Although self rated measures for depression exist, they do not
appear to have been successful thus far in attempts to differentiate
endogenous fron non-endogenous depression. A major reason for this
failure is undoubtedly due to the fact that the scales were developed
as measures of severity, and not intended or designed to differentiate
among subtypes. A second reason is the limited range of symptoms
covered by each instrument. Although there remains some question about
the validity of using patient‘s accounts of their symptomatology,
interest in developing a self rated scale remains strong due to
relative inexpense, ease of administration and their ability to

quantify feelings of the patient.

30

Table 2-2

Frequency of Citations for Symptoms
Associated with Endogenous Depression

 

 

Number of Literature

 

Symptom Citations*
1. Psychomotor Change 12
2. Autonomous Course 7
3. Self Reproach 7
4. Distinct Quality of Mood 7
5. Insomnia (Middle or Terminal) 6
6. Lack of Precipitants 3
7. Loss of Interest 3
8. Severity of Mood 3
9. Decreased Concentration 2
IO. Delusions 2
11. Diurnal Variation 2
12. Decreased Appetite 2
13. Family History 2
14. Premorbid Personality 2
15. Suicidal Ideation or Behavior 2
16. Pessimism 2
17. Body Type 1
18. Obsessions/Compulsions I
19. Impulse Inhibition 1
20. Loss of Insight 1
21. Social Hithdrawal 1
22. Sudden Onset 1

*Twenty-two principal investigations were reviewed.

 

CHAPTER 3
METHODOLOGY OF THE STUDY

In Chapter Three, the development of the Differential Diagnostic
Depression Scale (ODDS) is described. The sample, design, item
selection criteria, and the analytic procedures used in the study are

also presented.

Description _p_f_ the Development _o_f__t_h_e_ ODDS

 

 

A review of the relevant research, presented in Chapter Two, was
conducted by the author to identify symptom patterns which had been
associated with depression. Particular interest was paid to those
symptoms that were believed to differentiate endogenous and
non-endogenous depression. A research team consisting of a Ph.D.
counseling psychologist and eight doctoral level counseling psychology
students generated a pool of 141 items designed to measure the symptoms
identified in the literature.

The process of item generation, editing, and selection was
designed to maximize content validity. Only those items which were
consensually validated by all members of the research team were
included in the final item pool. As an additional check to ensure the
content validity of the items, the questionnaire was reviewed by a
psychiatrist who had considerable clinical and research experience in
the field of the diagnosis and treatment of depression. The result of

this collaborative effort was the Differential Diagnostic Depression

31

32

Scale (0005) which was designed to measure, through self report,
depressive symptomatology (Farquhar, Holmes, and Azar, 1982). A
summary of the symptoms and number of associated items may be found in
Table 3-1.

In addition to items from the twenty-nine symptom categories,
twenty items from a modified version of the Crowne-Marlowe Social
Desirability Scale (Crowne and Marlowe, 1964) were included in the
construction of the 0005. These items were included to permit future
research of the relationship between response to the symptom category
items and social desirability. Twenty items were randomly selected
from the Modified Crowne-Marlowe Scale and included as every eighth
iten on the 0005. The fOTlowing are examples of such items: "I can
remember playing sick to get out of something," and “I am careful about
my manner of dress.“

The ODDS had 161 items in the completed form, representing the
twenty-nine symptom categories listed in Table 3-1 and the items from
the Modified Crowne-Marlowe Social Desirability Scale. The distribu-
tion of items per symptom category is also listed in Table 3-1. The
141 symptom items were randomly ordered for presentation in the
questionnaire» 'The subjects were asked to respond to each item on a
four point Likert-type scale ranging from “Definitely True of Me"
through “Definitely Not True of Me." The complete questionnaire is
attached as Appendix A. The first page of the 0005 consisted of a fact
sheet which was used to gather information about demographics of the

sample .

Table 3-1

Symptom Categories
Represented on the 0005
m
Category Number of Items

Psychomotor Change 10
Autonomous Course

Self Reproach

Distinct Quality of Mood
Insomnia (Middle or Terminal)
Lack of Precipitants

Loss of Interest

Severity

Decreased Concentration
Delusions/Hallucinations
Diurnal Variation

Decreased Appetite

Family History

Suicidal Ideation or Behavior
Pessimism/Hopelessness
Obsessions/Compulsions
Impulse Inhibition

Social Hithdrawal

Sudden Onset

20. Anxiety

Crying

22. Memory Loss

Responsibility

Somatic Complaints

Length of Illness

26. Irritability

Fear

Sadness

Indecisiveness

magma-bump
O

HHH
NHO
O O O O

HHHO—‘HHH

AOQNO‘U'lbw
O O O O O O
0-.

N
g—I
o

N
w
o

N
b
o

0-.
NwwNHmuwmwwomwhmc-wmmbmmmast-Now

N N
\J U“
o o

NN
90m
0

 

34

Sample
The sample of the study consisted of two groups: individuals with

a diagnosis of probable or definite endogenous depression (n=50) and
individuals with a diagnosis of non-endogenous depression (n=50).
Sample selection criteria and diagnostic procedures for the two groups
is presented in detail in the next section.

The demographic information for both groups is presented in Tables
3-2 through 3-9. The mean age of the non-endogenous group and
endogenous group was 35.54 and 41.82, respectively. The majority of
both depressed groups were Caucasian with a small percentage of Asians
and Blacks (see Table 3-3). A significantly greater proportion of the
endogenous group were female (p<.05, see Table 3-4). The majority of
both groups were married, had been married only once, and had between
one and two children in their families (see Tables 3-5 to 3-7). Both
groups ranged in income from below $4,000 to over $40,000, and the
majority of both groups had at least high school educations (see Tables
3-8 and 3-9). ‘

Procedures

 

Subjects in the study were recruited following their admission to
the inpatient psychiatric unit at Ingham Medical Center, Lansing,
Michigan, or after their intake at a consortium of Lansing area out-
patient psychiatric clinics. The outpatient clinics included the
Affective Disorders Clintc at Michigan State University; Eaton
Counseling Center, Charlotte, Michigan; and the Ingham County Mental
Health Center, Lansing, Michigan.

35

Table 3-2
Age of Respondents

 

 

 

 

 

 

 

Endogenous Non-Endogenous
Mean 41.82 35.54
Median 39 33
Standard Deviation 6.45 5.93
Table 3-3
Race of Respondents
Endogenous Non-Endogenous
Race Percent Percent
Asian 4 0
Black 4 4
Caucasian 88 92
Other 4 4
Table 3-4
Sex of Respondents*
Endogenous Non-Endogenous
Sex Percent Percent
Female 76 56
Male 22 44
No Response 2 0

*p<.05

 

36

Table 3-5

Marital Status of ReSpondents

 

 

 

Endogenous Non-Endogenous
Category Percent Percent
Single 28 26
Married 40 46
Living Together 8 4
Hidowed 2 o
Divorced 16 20
Separated 2 2
No Response 4 2

Table 3-6

Respondent's Number of Marriages

 

 

 

Endogenous Non-Endogenous
Category Percent Percent
None 38 42
One 34 48
Two 24 10
Three 2 0
No Response 2 0
Table 3-7

Respondent's Number 5 Age of Children

 

 

Endogenous Non-Endogenous
Category Mean Mean
Number 1.7 1.2
Age of Youngest Child 7.22 5.98

Age of Oldest Child 11.84 7.64

 

37

Table 3-8

Income of Respondents

 

 

Endogenous Non-Endogenous
Category Percent Percent
Under 34,000 18 20
$4,000-36,000 16 4
36,000-310,000 6 6
310,000-315,000 10 I6
315,000-320,000 8 10
320,000-325,000 14 14
325,000-330,000 10 14
330,000-340,000 B 6
Over 340,000 6 6
No Response 4 2

Table 3-9

Education of Respondents

 

 

Endpgenous Non-Endogenous
Category Percent Percent
Grade School , 4 0
Junior High 0 2
High School 58 36
Trade School 12 10
Associate's Degree 10 O
Bachelor's Degree 6 18
Master's Degree 8 22
Doctoral Degree 0 8
No Response 2 4

 

38

Inclusion Criteria

The subjects were selected for the study if, during the initial
screening or intake, they presented with subjective complaints of
depression, dysphoria, or experienced a loss of interest or pleasure in
their usual activities. Patients did not qualify for inclusion in the
study if depression was suspected by the cflinician but denied by the
patient. The subjects were between the ages of 18 and 65 and free of
medical illnesses that may have invalidated the results of their
dexamethasone suppression test described below. Subjects were also
excluded if their depression was secondary to other psychiatric
illnesses. The medical and psychiatric exclusions criteria are listed
in Figure 3-1.

During the admission or intake procedure, or as soon as it was
deemed clinically appropriate by the hospital or clinic staff, those
patients who qualified for inclusion were informed by the staff about
the nature of the study and asked if they would be willing to partici-
pate. At that time, they were told that their decision to participate
would not influence the treatment that they would receive. If they
were willing to participate, a consent form was obtained for each
subject (see Appendix B).

After subjects signed their informed consent to participate, three
procedures were followed. First, each subject was asked to complete
the 0005. After the completion of the ODDS, each subject was given a
semi-structured interview, the Schedule for Affective Disorders and
Schizophrenia (Endicott and Spitzer, 1978). The diagnosis of endoge-
nous or non-endogenous depression was made based on the results of the

interview, using Research Diagnostic Criteria (Spitzer and Endicott,

39

Figure 3-1

Psychiatric and Medical Exclusion Criteria

 

 

 

Psychiatric

1. Schizophrenia

2. Bipolar Depression

3. Organic Brain Syndrome

4. Alcoholism

5. Anorexia Nervosa

Medical*
1. Pregnancy, high dose estrogen therapy other than oral
contraceptives.

2. Cushing's disease or syndrome

3. Severe weight loss where body weight (80% of ideal weight
4. Hepatic enzyme induction (phenytoin sodium, barbituates,

mepro-bamate)

5. Uncontrolled diabetes mellitus (hypoglycemia, acidosis)

6. Major physical illness; trauma; fever; dehydration; nausea
7. Temporal lobe epilepsy; use of reserpine or narcotics

8. Addison's disease

9. Corticosteroid therapy
10. Hypopituitarism

11. High dose benzodiazepines ()25mg/day of diazepam)

12. Other endocrine disease

13. Spironolactone therapy

*Adapted from B.J. Carroll, et. al. A Specific Laboratory Test

for the Diagnosis of Melancholia. Archives of General
Psychiatry, 1981, 38, 15-22.

 

40

1978). Finally, each subject was given the Dexamethasone Suppression
Test (DST), a laboratory test for blood plasma cortisol level. If the
subject was on an inpatient unit, 1 mg. of dexamethasone was
adninistered at 11:00 p.m. by unit staff. Blood samples were then
drawn by hospital staff at 4:00 p.m. the following day. If the subject
was an outpatient, they were given the dexamethasone for self admini-
stration and instructed to return to the clinic or a designated
laboratory for the blood sampling the following day at 4:00 p.m.
Although Carroll (1982) reports a higher sensitivity with the DST if
blood is drawn more frequently, the 4:00 p.m. blood draw was used for

practical considerations.

Rater Training

 

The interviewer/raters for this study consisted of seven doctoral
students in a psychology program who were part of a research seminar on
depression. The training materials and two stage procedure used in
the study were drawn from a training program specifically developed for
the RDC by Gibbon and her colleagues (1981). First, the raters were
asked to read the RDC as well as articles which described development
and clinical use of the instrument (Endicott and Spitzer, 1978;
Spitzer, Endicott, Robins, 1978). The raters discussed the instruments
and resolved differences in interpretation of the diagnostic criteria.

In the second stage of training, 14 written case vignettes were
rated by each participant to gain experience in using the RDC. The
vignettes, abstracted from actual patient records, included chief
complaint, history of present illness, a review of past psychiatric
history, and description of findings from a mental status examination.

The raters independently rated the case vignettes. After

41

completing each vignette they compared their ratings with a key. The
keys consisted of consensus ratings by three experienced raters. The
keys were supplied with the vignettes by the developers of the training
program (Gibbon, et. al. 1981). For this study, each rater' was
encouraged to discuss the rationale of the rating of the vignettes in a
group format.

Following the training procedures each person rated seven
additional written case vignettes using the RDC. In addition, raters
were asked to rate two patient videotaped interviews using the RDC
criteria.

A two—way analysis of variance was used to compare the responses
of the raters on each diagnostic category across the nine test cases.
An_ intraclass correlation coefficient of .97 was obtained for major

depressive disorder, endogenous subtype.

Classification Procedures

 

For the purpose of analysis, depressed subjects were designated as
endogenous or non-endogenous according to the flow chart presented as
Figure 3-2. Subjects with a positive, or abnormal Dexamethasone
Suppression Test (>5mg/dl) were classified as endogenous depressions.
Those with negative or normal (55mg/dl) were classified according to
their RDC diagnosis. The Dexamethasone Suppression Test has been
demonstrated to be a highly specific biological marker for endogenous
depression, with a specificity rate of 96% (Carroll, et. al., 1981).
Individuals with a research diagnosis of "Major Depressive Disorder,
Probable or Definite Endogenous" subtype were classified as endogenous
depressions. .All other research diagnoses for depression were

classified as non-endogenous depressions.

42

Figure 3-2
Diagnostic Classification of Patients

   

    
     
     

Doee the
enhjeet neet
not for other
depreeainni

   
     
   
 
      
   

  
  

Upon intake oi
aeiaeion. doee the client

patient oeeeent ieeiinee and/or

  

eyeotoea ei
deeeeeaieni

yea no

i

~Adnigia§lggggns
i. Ieeeaeeh Questionnaire

 

Inc (or eejoe
lapeeeeive

lieotder?
IUDCYFF
lndoaenoua?

2. Dean-e t haeone
Suppression feet (081)

3. sane/Inc

 

 

 

yea yea yea

eooi oi Endogenous-
Ielanehoiie eebjeeta

 

 

 

 

 

 

 

pool oi Moe—ledoaeeoee
eehjecte

 

the 0005 were formed by two methods of item selection: an “a priori"

method and an "empirical" method.

The design for the study was essentially correlational.

43

m

constructed by grouping items with similar content together.

were then eliminated from each scale if they decreased the internal
consistency of the scale.
grouped into scales based on their highest loading on factors extracted
by principal components analysis, which is correlational in nature.

Both methods of item selection were compared by computation of a

discriminant function, which is also correlational in nature.

3'3e

level to ensure that the endorsement rate for each item was within an
acceptable range.

.95 were to be withheld from further analyses.

Research Hypotheses

The main hypotheses for the study were as follows:

1.

2.

A flow chart of the analytic procedures is presented as Figure

The first activity consisted of examining the item endorsement

A self report instrument formed by the a priori method
will differentiate between endogenous and non-endogenous
depression.

A self report instrument constructed by the empirical
method will differentiate between endogenous and
non-endogenous depression.

Differences between endogenous and non-endogenous
depression found through hypotheses #1 and #2 will lead
to interpretable constructs about the nature of endoge-
nous and non-endogenous depression.

Analysis Procedures

 

Scales on

In the a priori method, scales were

In the empirical method, items were first

Items with an endorsement level below .05 or above

The rationale for

 

 

Ahcve .15

44

Figure 3-3

Analytic Procedures For DDDS.)

 

 

 

Exp-inc
Responses 1222£££

(ltee Frequencies)

    

  
   
   
 

ltee Frequency
8.95. (.05

 

 

 

 

 

 

    
    

 
 

ggnsigcr

is: F scale

 

 

 

 

A Priori
Approach

 

 

 

 

 

 

    
 

 

 

 

 

 

   
 

  
 
   
   

 

 

 

 

 

 

 

 

  

  
   
    

22222 1:... er
r-——‘L-—l A priori Classifi-
Discard cetion into Scales
£5355 Internal
Consistency
values of Scales
“ Internal
3.1.9.3.! ” l_ V" Consistency highs
iron original 6?
Scale [7 r ital delete
)
E
Chi-Square
Analysis
humus
Significance ..l°' '15

level

 

Place item
in E Scale

 

 

 

 

 

 

lapiricsl
Approach

 

 

 

 

 

Correlation
Matrix

 

Place its-s into ‘ ®

 

 

 

pivgde ss-ple into

Endogenous and Mon- __.®

Endogenous groups

 

 

45

Figure 3-3 (cont ' d)

    
  
 
  
   
 

 

 

 

23 Principal letsin Factors
I Cn-ponent with vith eigen values 2%D::::::'
one (I) in Diagonal greater than one
Lover ii itec
included

 

 

   

ns ett
Internal
Consistency
of Factors

  
  
 
 
 
   

 
  

 

E Discrisinant
Analysis

 

Higher if its. included
or r 3.39

 

 

 
 

131255; function
[or interpretation
and separation

 
   
 

46

withholding these items was that they would have little variance and
would not contribute to the discrimination of types of depression.

After screening all items for endorsement level, two alternate
strategies were followed for selecting items to be used in the
construction of scales. In the first approach items were grouped into
scales based on an a priori examination of item content. After this
grouping, the internal consistency of each of the scales was calculated
by the use of coefficient alpha. Each item in a given scale was
examined to determine whether the internal consistency would increase
if the item was deleted. Items that lowered the internal consistency
of the scale were removed from that scale. A three by four chi square
was used to analyze each item. Respondents were separated into three
diagnostic groups based on their RDC diagnosis of probable, definite,
or nonendogenous depression. Items that lowered the internal consis-
tency of a given scale but were significant at the .15 level or below
in the chi square analysis were removed from the original scale and
placed in a general scale labeled as endogenous, or E Scale.

In the second approach to scale construction, all items were
placed in a correlation matrix and a principal factor solution was done
without comnunalities in the diagonals. The principal factor solution
was used to examine the relationship among items and to find out how
the item responses related to one another. A varimax rotation was used
following the principal factor solution. This procedure maximizes the
within factor loading for each item. Only those factors with a sum of
squares (eigenvalue) in excess of one were rotated. Items that did not
load higher than .30 on any of the rotated factors were withheld from

further analyses. The internal consistency of each factor was computed

47

with coefficient alpha, and items that lowered the reliability of a
given factor and did not load higher than .39 were withheld from
further analyses. However, some of these items were retained as part
of the scale in an attempt to create an equal number of items as the a
priori approach.

After the research scales were constructed by the two approaches,
the scales were used in a discriminant analysis in an attempt to find a
combination of scales that would best discriminate between the two
groups. Thirty-five subjects from each diagnostic group were randomly
selected for the discriminant analysis and fifteen subjects from each
group were withheld for puroses of cross validation. The discriminant
analysis was used to compare actual versus predicted group membership
for each of the two diagnostic categories. The percentage of correct
classification for known groups was compared for the a priori and the
empirical method. The discriminant function derived from the analysis
phase was used for both methods to predict the group membership for
those subjects withheld for cross validation.

A second discriminant function was constructed for each approach
using those scales identified in the literature as the best predictors
of endogenous depression. The discriminant analysis was conducted on
the seventy subjects identified above, and cross validated on the group
of thirty individuals withheld for that purpose.

Finally, the individual items that discriminated between the
diagnostic groups placed in a separate scale. The group means on the
scale were compared for the seventy randomly selected subjects. A
second comparison of group means was conducted for the two cross

validation groups.

48

Sunlna ry

The literature on endogenous and non-endogenous depression was
examined for symptom categories of each type of depression. A research
team generated 141 items that were believed to match the content of
those categories. The items were conbined with twenty items from the
Modified Crowne-Marlowe Scale, and the resultant instrument was named
the Differential Diagnostic-Depression Scale (ODDS). A sample of 100
depressed individuals was obtained through a consortium of inpatient
and outpatient psychiatric centers. The ODDS was administered to each
subject as well as the Dexamethasone Suppression Test (DST) and the
Schedule for Affective Disorders and Schizophrenia (SADS). Individuals
were diagnosed as endogenous or non-endogenous depression based on a
combination of Dexamethasone Suppression Test values and Research
Diagnostic Criteria (RDC).

In the data analysis, two approaches were used to construct scales
from individual items. In the a priori approach, items were grouped in
scales prior to analysis based upon similarity in content. The
internal consistency for each scale was calculated, and items were
deleted if they lowered the reliability Of the scale. A chi square
analysis was done on all of the 0005 items. Items that discriminated
at the .05 level or below were retained in a general scale. In the
second approach to the construction of scales, a principal conponent
solution was used on all items. Factors were eitracted until the sum
of squares (eigenvalue) was less than 1.0 or the factor was not
interpretable. The internal consistency of each\9f the factors was

determined, and items were deleted that lowered the reliability and had

49

factor loadings of less than .39. Some of the a priori scales were
omitted to ensure an equal number of items in each approach.

After the two methods were used to construct two separate sets of
scales, a discriminant analysis was used to find a discriminant
function which would accurately classify the depressed groups. The
accuracy of each method was calculated to compare the efficiency of the
two procedures. A second discriminant analysis was conducted using a
select number of scales for each approach. Finally, individual items
that discriminated between endogenous and nonendogenous items were

placed in a scale and a comparison of group means was conducted.

CHAPTER FOUR

Results of the Data Analysis

In Chapter Four, the results of the tests of each of the criteria
for item selection as well as the major hypotheses are presented. The
results of the tests for item selection for both the a priori and
empirical method are presented in the first section, with emphasis
placed on the items that were retained or withheld from further
analysis. The results from the discriminant analyses for both

approaches are presented in the second section.

Item Endorsement Level

 

The distribution of responses for each item was examined to
determine the item endorsement level. The rationale behind this
examination was that items that were endorsed in one direction by a
large proportion of the sample would have little variance. None of the
responses to the individual items had an endorsement level that
exceeded .95. All items, with the exception of the social desirability
items, were retained for further analysis. The social desirability

items were held for future research efforts.

A Priori Scale Construction

 

In the a priori method of item selection, items were grouped in
scales according to content similarity (See Table 3-1). The internal
consistency of each scale was computed using Cronbach's alpha

SO

51

coefficient. If the internal consistency increased when an item was
eliminated it was removed from the scale. All items were also screened
by chi-square analysis. Items that lowered a scale internal
consistency but were significant at the .15 level or below were removed
from the original scale and placed in a heterogenous scale labeled
endogenous or E Scale.

Fifteen items decreased the internal consistency of the original
scales and were not signficant at the .15 level or below. These items
are listed in Appendix C with the respective alpha coefficients of
their original scale. Ten additional items were found to lower the
internal consistency of their original scale if included and were
significant at the .15 level or below. These items were removed from
their original scale and placed in the E Scale and are listed in
Appendix D. .

Following the examination of the internal consistency and
chi-square tests, 12 additional items were withheld from further
analysis in order to match the number of items generated by the
empirical approach to item selection. This was done to ensure an equal
number of items for both approaches for the discriminant analysis. The
12 items were nested in four scales and are listed in Appendix E. The
criteria for the elimination of these items was that: 1) none of the
items were significant at the .15 level or below; and, 2) none of the
four scales (sadness, fear, indecision, and social withdrawal) had
prior research support as good discriminators of subtypes of
depression.

After the 27 items were deleted, 114 items remained for further

analysis. The retained items are listed by their respective scale in

52

Table 4-1. Alpha levels are included for 21 of the 25 a priori scales.
The reliabilities ranged from .57 to .88. The alpha level was not
computed for four of the scales due to the SPSS program limitation that
each scale have at least three items. A correlation matrix for the 25

scales is presented as Appendix F.

Empirical Scale Construction

In the empirical approach of scale construction, all 0005 items
except for the social desirability items were placed in a correlation
matrix. A principal factor solution was done without communalities in
the diagonals. A varimax rotation was executed for factors with an
eigenvalue greater than one.

Fifteen interpretable factors were extracted from the above
analysis as follows: Factor 1: General; Factor 2: Self Reproach; Factor
3: Suicidal Ideation and Behavior; Factor 4: Agitation; Factor 5:
Anorexia and Height Loss; Factor 6: Dependency; Factor 7: Interpersonal
Dissatisfaction; Factor 8: Memory Impairment; Factor 9: Crying; Factor
10: Body Integrity; Factor 11: Delusions and Hallucinations; Factor 12:
Pessimism; Factor 13: Loss of Energy; Factor 14: History; Factor 15:
Diurnal Variation and Somatic Anxiety. Eighteen of the items had low
loadings on the 15 factors (below .30) and were withheld from further
analysis.

The internal consistency for the retained items on each factor was
computed using the Cronbach alpha coefficient. Nine items were found
to lower the internal consistency of a factor and had factor loadings
less than .39. These nine items were withheld from further analysis
and can be found in Appendix G. The internal consistency of the

factors ranged between .31 to .92.

53

Table 4-1

Scales Determined by
the A Priori Approach

 

SCALE 1
Item #
86
100
153

Scale 2
Item #
31

51

89

109

Scale 3
Item #
37

39

131

Scale 4
Item #
19

52

103

138
156
161

Scale 5
Item i

60

Scale 6
Item i
11

35

43

75

Scale 7
Item #
15

20
43

ANXIETY; Reliability ' .75

As the days pass, I seem to get more nervous.
Sometimes I get so nervous, I panic.
I feel tense.

CRYING; Reliability = .82

I continue to cry over the event that caused my depression.
I seem to cry for no reason at all.

I'm really tired of crying.

I can't cry even when I feel like it.

CONCENTRATION; Reliability 3 .83

I can't seem to keep my mind on one thing.
I can't concentrate.
I can hardly think at all.

DELUSIONS/HALLUCINATIONS; Reliability 8 .71

I hear things that other people don't hear.

I feel as if things are not real.

I have felt that people have been making me do things by
controlling my mind.

I think of strange things that are too bad to talk about.

I see things that other people don't see.

I hear strange things when I'm alone.

EATING DISTURBANCE OR HEIGHT LOSS; Reliability = .77

Food doesn't taste good to me anymore.
I've been losing weight recently, even though I'm not trying
to.

IMPULSE INHIBITION; Reliability 8 .78

I hold myself back from doing what I want.

I can't let go and have fun anymore

I feel so bad I don't feel anything anymore.

I can't remember the last time I had a good laugh.
I don't allow myself to get involved.

REACTIVITY; Reliability - .66
I feel my condition is hopeless.

Nothing seems to make me feel better.
I feel so bad I don't feel anything anymore.

Scale 8

em #

7

58

99

102

119

162

Scale 9
Item 5

Scale 10

Item #

5
68
137
150

Scale 11

Scale 13

Item #
30
101

54

Table 4-1 (continued)
LOSS OF INTEREST; Reliability I .76

I have little interest in being with friends.

I've lost interest in doing my work.

I've lost my interest in sex.

I have little desire to eat.

I eat less now than usual.

Most of the time, I'm not interested in sex, but once in
awhile I am.

MEMORY; Reliability . .88

I can't count on my memory.
I have a hard time remembering things.
I don't remember things as well as other pe0ple do.

OIURNAL VARIATION; Reliability 8 .71

I wake up refreshed.

I feel worse in the morning than any other time in.the day.
I wake up depressed and feel like things are going to get
worse.

I feel tired and jittery when I wake up.

ONSET; Reliability 8 .51

My depression came on me quickly.
My depression seemed to come out of nowhere.

HOPELESSNESS; Reliability 8 .85

I have no hope for myself.
Things aren't going to get any better for me.
No one can help me now.

HISTORY; Reliability 8 .74

Other people in my family have had problems with depression.
I mm the only one in my family that I know of that has had
problems with depression.

Someone in my family has been hospitalized for depression.

PSYCHOMOTOR; Reliability I .79

I feel like my heart is racing.

I don't have any energy.

I'm so tired I don't care about anything anymore.

I feel nervous and edgy.

Everything bout me is slowed down.

I'm tired most of the time, but sometimes I have a burst of
energy.

I feel like my mind is racing.

Item 4
87
124
I45

'Spale 15
Item #
77

82

93

94

95
110

Scale 16

Scale 17

Item 4

26
27
28
47
62
73
76

Scale 18
tem

21

81

85

108

134

Scale I9
‘_—'36

78
114

Scale 20
Item #
10

I4

17

18

59

66

92

55

Table 4-1 (continued

I'm ashamed of the fact that I feel so tired.
I feel like my body is speeded up.
I've come to live with the fact that I'm always tired.

RESPONSIBILITY; Reliability c .73

My ability to manage my life changes alot.

I have lost my sense of responsibility.

People can't count on me anymore.

I could not survive unless someone took care of things for
me.

Although I feel bad, I can do things if I have to.

I cannot manage my own life.

OBSESSIVENESS; Reliability 3 .57

I have the same thoughts over and over again.
I can't seem to get rid of thoughts that bother me.

SELF REPROACH; Reliability - .85

feel worthless and no good.

can't stand myself anymore.

don't like myself.

feel empty and hollow.

don't see how anyone can stand being around me.
feel powerless.

feel guilty.

HHHHHHH

SEVERITY; Reliability = .75

My depression is so overwhelming, I find it hard to go on.
My depression is so severe I drive people away.

I'm very depressed.

I am more depressed than anyone I know.

My depression varies from mild to severe.

INSOMNIA; Reliability . .81

I wake up early in the morning and can't back to sleep.
I have problems going to sleep.

I wake up alot during the night.

SOMATIC; Reliability e .82

My chest feels tight.

I have problems with constipation.
People tell me I sigh alot.

I have problems with diarrhea.

My body feels like it is rotting inside.
I have problems with headaches.

I have chest pains frequently.

Item i
107

129
132
133
135
143
149

Scale 21
Item #
33

42

98
146

Scale 22
em #
57

Scale 23
Item #
225

Scale 24
Item #
115

111

Scale 25
Item #
I48

71

56

Table 4-1 (continued)

Although people don't believe it, I know there is something
wrong with my body.

I have a lot of hard to place aches and pains.

I have difficulties breathing.

My stomach seems to be upset alot.

My body is going to pieces.

My breathing seems different now.

I find myself sighing alot.

SUICIDE; Reliability - .84

I've made a serious attempt to harm myself.

I tried to comit suicide, but I knew it wasn't going to
I95: tried to commit suicide.

I think about suicide alot, but I know I won't do it.
DISTINCT 0UALITY*

I'm more depressed than usual.

LENGTH*

My depression started less than a year ago.

IRRITATION*

I get upset easily.
The smallest things seem to upset me.

PRECIPITANT*

I know what caused my depression.
I get depressed over nothing.

*Reliability not calculated as the original scale had less than 3 items

57

After the 27 items were deleted in the empirical approach, 114
items remained for further analysis. These items are listed along with
their respective factor loadings in Table 4-2. .A correlation matrix
for the 15 factors is presented as Appendix H. The correlation matrix

was computed for the retained items on each factor.

Discriminant Analysis

 

Following the construction of the research scales by the a priori
and empirical approach, a discriminant analysis was exetuted to find a
combination of scales that would best descriminate between the two
depressed groups. Thirty-five subjects from each diagnostic group
(endogenous and non-endogenous) were randomly selected for the
analysis, and the remaining subjects were held for cross validation.

One discriminant function was constructed for each approach. The
standardized function coefficients for the variables used in both
approaches is presented in Table 4-3. Each coefficient indicates the
relative contribution of the variable to the discriminant function. In
the a priori approach, the scales anxiety and precipitant made the
largest relative contribution to the discriminant function, whereas the
scales impulse inhibition and length made the smallest contribution.
In the empirical approach, the scales pessimism and general made the
largest contribution, whereas suicidal ideation and crying contributed
the smallest amount to the discriminant function.

The discriminant functions were then used to classify the seventy
cases that were randomly selected from the two depressed groups. The
classification results are presented in Tables 4-4 and 4-5. Eighty
seven percent of the seventy cases were classified correctly with the a

priori approach versus 77% correct classification with the empirical

58

Table 4-2
Scales Selected by the Empirical Approach
._____________________________________________________________________________

Factor 1: GENERAL; Reliability = .92

Item # Factor Loading
44. Everything about me is slowed down. .78
43. I feel so bad I don't feel anything anymore. .73
158. All the joy is gone out of my life. .71
47. I feel empty and hollow. .69

54. I can't remember the last time I had a good laugh. .67

57. I'm more depressed than usual. .63
45. I don't take chances anymore. .58
35. I can't let go and have fun anymore. .58
75. I don't allow myself to get involved. .56
85. I'm very depressed. .55
131. I can hardly think at all. .52
97. I feel sad. .51
20. Nothing seems to make me feel better. .49
150. I feel tired and jittery when I wake up. .47
73. I feel powerless. .45
139. I can't seem to think about anything but how
bad I feel. .45
2. I don't have any energy. .44
130. I want to be alone most of the time. .42
59. My body feels like its rotting inside. .41

147. I usually get along with people but lately my
seem to have been falling apart. .39

58. I've lost interest in doing my work. .39

59

Table 4-2 (continued)
Item # Factor Loading

142. I'm afraid about what will happen in the future. .38

69. I have pulled into myself. .37
21. My depression is so overwhelming, I find it hard

to go on. .35
93. People can't count on my anymore. .33

29. I have been slowly getting more and more
depressed over the past few weeks. -.47

Factor 2: SELF REPROACH/ANXIETY, Reliability 8 .9O

26. I feel worthless and no good. .71
28. I don't like myself. .67
27. I can't stand myself anymore. .61
12. I watch what I am doing with other people. .58
11. I hold myself back from doing what I want. .58
6. I have many different fears about the present

and the future. .54
76. I feel guilty. .52
41. I don't seem to have any control over my life. .47
36. I feel nervous and edgy. .45
86. As the days pass, I seem to get more nervous. .44
74. I have the same thoughts over and over again. .36
100. Sometimes I get so nervous, I panic. .32

Factor 3: SUICIDAL IDEATION AND BEHAVIOR, Reliability - .86
98. I've tried to commit suicide. .82
33. I've made a serious attempt to harm myself. .78

42. I tried to commit suicide, but I knew it wasn't
going to work. .77

Item #
63.
146.

53.

117.
121.

Factor 4:
123.

124.

1.

IO.

37.

78.

Factor 5:
102.
119.
60.

9.
157.
38.

60

Table 4-2 (continued)

I have been hospitalized for depression.

I think about suicide a lot, but I know I won't

do it.

There have been times in my life when I had no

idea what I did.
I've done things too terrible to think about.

If I weren't so afraid of dying, I'd commit
suicide.

AGITATION, Reliability = .82

I feel like my mind is racing.

I feel like my body is speeded up.

I feel like my heart is racing.

My chest feels tight.

I can't seem to keep my mind on one thing.

I have problems going to sleep.

ANOREXIA/HEIGHT LOSS, Reliability = .86
I have little desire to eat.
I eat less now than usual.

I've been losing weight recently, even though
I'm not trying to.

Food doesn't taste good to me anymore.
I eat more when I get depressed.

I wake up early in the morning and can't
get back to sleep.

Loading
.64

.54

Item i
Factor 6:

94.

110.
95.

81.
82.
127.
39.
99.

Factor 7:

126 O

111.
62.
65.

105.
83.

115.

162.

Factor 8:
91.
67.

61

Table 4-2 (continued)

Factor

DEPENDENCY, Reliability 8 .82

I could not survive unless someone took care
of things for me.

I cannot manage my own life.

Although I feel bad, I can do things if I have to.
I have little interest in being with friends.

My depression is so severe I drive people away.

I have lost my sense of responsibility.

I find it impossible to make a decision.

I can't concentrate.

I've lost my interest in sex.

INTERPERSONAL DISSATISFACTION, Reliability 8 .83

I have had difficulties getting close to people
for a long time.

The smallest things seem to upset me.

I don't see how anybody can stand being around me.
I pull away from people.

I seem to drive people away.

I get irritated a lot.

I get upset easily.

Most of the time I'm not interested in sex,
but once in a while I am.

MEMORY IMPAIRMENT, Reliability 8 .62
I have a hard time remembering things.

I can't count on my memory.

Loading

.81
.78

Item #
118.

84.

Factor 9:
51.
109.
89.
66.
31.

Factor 10:

135.
133.

92.
107.

129.
159.

Factor 11:

19.

161.
103.

138.

62

Table 4-2 (continued)
Factor Loading

I don't remember things as well as other people
seem to. .71

I can't do anymore than I'm doing now. .34

CRYING, Reliability 8 .80

I seem to cry for no reason at all. .72
I can't cry even when I feel like it. .63
I'm really tired of crying. .61
I have problems with headaches. .58

I continue to cry over the event that caused .
my depression. .52

BODY INTEGRITY, Reliability 8 .58

My body is going to pieces. .57
My stomach seems to be upset a lot. .53
I have chest pains frequently. .47
Although people don't believe it, I know

there is something wrong with my body. .41
I have a lot of hard to place aches and pains. .40
My body seems to be working all right. -.69

DELUSIONS/HALLUCINATIONS, Reliability 8 .67
I hear things that other people don't hear. .74
I hear strange things when I'm alone. .74

I have felt that people have been making me do
things by controlling my mind. .48

I think of strange things that are too bad
to talk about. .42

63

Table 4-2 (continued)

Item # Factor Loading
Factor 12: PESSIMISM/HOPELESSNESS, Reliability 8 .31

49. Things aren't going to get any better for me. .54
149. I find myself sighing a lot. .50
156. I see things that other people don't see. .48

34. I have no hope for myself. .39
148. I know what caused my depression. .38

155. I can't seem to get rid of some of the thoughts
that bother me. -.60

113. When things are going well I feel like I have
more energy. . -.68

Factor 13: FATIGUE, Reliability 8 .72

50. I'm tired most of the time, but sometimes I have

a burst of energy. .65
145. I've come to live with the fact that I'm

always tired. .61
87. I'm ashamed of the fact I feel so tired. .61

Factor 14: HISTORY, Reliability 8 .72

101. I am the only one in my family that I know of
who has problems with depression. .80

30. Other people in my family have had problems
with depression. .78

122. Someone in my family has been hospitalized
for depression. .69

64

Table 4-2 (continued)

Item i Factor Loading

Factor 15: OIURNAL VARIATION/SOMATIC ANXIETY, Reliability 8 .52

68. I feel worse in the morning than any other time

in the day. .61

17. People tell me I sigh alot. .60

154. I feel miserable in the morning. .48
132. I have difficulties breathing. .36
55. I'm afraid of being alone. .35

I4. I have problems with constipation. .34

65

Table 4-3

Standardized Discriminant Function Coefficients
for the A Priori and Empirical Approaches

 

 

 

A PRIORI EMEIBlCAL
Variable Coefficient Variable Coefficient
Anxiety -.901 General -.601
Crying -.182 Self Reproach -.333
Concentration -.616 Suicidal Ideation
Delusions/Hallucinations .204 and Behavior -.O38
Eating .130 Agitation .142
Impulse Inhibition -.020 Anorexia/Height Loss .5237
Reactivity -.O91 Dependency .296
Loss of Interest .379 Interpersonal Dissatisfaction .337
Memory .317 Memory Impairment .359
Diurnal Variation .602 Crying -.078
Onset -.108 Body Integrity .230
Pessimism .343 Delusions/Hallucinations .101
History .112 Pessimism .641
Psychomotor Change .324 Fatigue .050
Responsibility .078 History .043
Obsessiveness .129 Diurnal Variation]
Reproach -.102 Somatic Anxiety .373
Severity .286 A
Sleep .493
Somatic .249
E Scale .314
Distinct Quality -.356
Length -.028
Irritation .299
Precipitant .618
Suicide -.461

 

 

66

Table 4-4

Classification Results of Discriminant Function
with A Priori Scale Construction

 

 

 

 

 

Number Predicted Group Membership
Actual Group of Cases Endogenous Non-Endogenous
Endogenous 35 29 6
Non-Endogenous 35 4 31
2 -
X " 39e4, p<e05
Table 4-5
Classification Results of Discriminant Function
with Empirical Scale Construction
Number Predicted Group Membership
Actual Group of Cases Endogenous Non-Endogenous
Endogenous 35 28 7
Non-Endogenous 35 9 26

x? = 33.11, p<.005

 

67

approach. The discriminant function for the a priori approach was
significant at the .05 level, whereas the discriminant function for the
empirical approach was significant at the .005 level.

Following the classification of the seventy cases, the discrimi-
nant function was used to cflassity the thirty subjects that had been
withheld for cross validation. Fifty seven percent of the cross
validation group were correctly classified by the discriminant function
constructed by the a priori approach (see Table 4-6). Forty three
percent of the cross validation group were correctly classified by the
discriminant function constructed by the empirical approach (see
Table 4-7).

A second discriminant function was constructed for each approach
using a reduced number of scales. The number of scales was reduced in
order to increase the ratio of subjects to scales. The scales were
selected for the new discriminant functions based on the review of the
literature in Chapter Two.

The following scales were used in the a priori approach: Eating,
reactivity, loss of interest, diurnal variation, psychomotor change,
self reproach, sleep, somatic, distinct quality of mood, and precipi-
tant. The discriminant function accurately classified 73% of the
original seventy subjects. In cross validation, the accuracy dropped
to 60%, which was not significantly different from chance classifi-
cation (p>.05).

The following factors were used in the empirical approach:
General, self reproach, agitation, anorexia/weight loss, body

integrity, diurnal variation/somatic anxiety.

68

Table 4-6

Cross Validation Results
of A Priori Discriminant Function

 

 

 

Number Predicted Group
Actual Group of Cases Endogenous Non-Endogenous
_
Endogenous 15 8 7
Non-Endogenous 15 6 9
Table 4-7

Cross Validation Results
of Empirical Discriminant Function

 

Number Predicted Group
Actual Group of Cases Endogenous Non-Endogenous
W
Endogenous 15 6 9

Non-Endogenous 15 8 7

 

69

The discriminant function accurately classified 73% of the
original seventy subjects and 57% of the cross validation group, which
was not significantly different than chance (p>.05).

Following the discriminant analysis, the eighteen items that
discriminated between probable, definite, and non-endogenous depression
via by chi-square analysis were placed in a scale. The group means
were compared for the endogenous and non-endogenous groups for the
seventy randomly selected subjects. There was a significant difference
between the two group means (p<.01). However, there was no significant

differences in the group means for the cross validation groups (p>.05).

Research Hypotheses

In Chapter Three, the following research hypotheses were stated:

1. A self report instrument formed by the a priori method will
differentiate between endogenous and non-endogenous depression.

2. A self report instrument constructed by the empirical method
will differentiate between endogenous and non-endogenous
depression.

3. Differences between endogenous and non-endogenous depression
found through hypotheses 1 and 2 will lead to interpretable
constructs about the nature of endogenous and non-endogenous
depression.

Based on the results of the original discriminant function with
the study sample, the first hypothesis was rejected. The classifica-
tion accuracy based on the a priori approach dropped from .87 to .57 in
cross validation, which is not significantly different than chance
classification. These results were confirmed with the second discrimi-

nant function using selected scales.

70

The second hypothesis was also rejected. The classification
accuracy based on the empirical approach dropped from .77 to .43, which
is not significantly different than chance. This was supported by the
second discriminant function using selected factors.

Due to the instability of the accuracy of classification for both
approaches, the third hypothesis was rejected. No clear, consistent

differences emerged that led to interpretable constructs.

Summary

The item endorsement level of all items in the 0005 was examined
to identify items that had an endorsement level greater than .95. None
of the items had an endorsement level greater than .95. All items,
with the exception of the social desirability items, were retained for
further analysis.

In the a priori scale construction items were grouped in twenty
nine scales according to content similarity. The internal consistency
of each scale was computed by coefficient alpha. The frequency
distribution of responses was analyzed by chi-square. Fifteen items
decreased the internal consistency of the scale and were not signifi-
cant at the .15 level or below. These items were withheld from further
analysis. Ten additional items lowered the internal consistency of the
original scale and were significant at the .15 level. These items were
placed in a general scale. One hundred and fourteen items remained for
further analysis. Twelve additional items were omitted to ensure equal
numbers of items in both approaches to scale construction.

In the empirical approach to scale construction, fifteen factors
were extracted by a principal component solution. The factors were

rotated by a varimax rotation. Eighteen items failed to load higher

71

than .29 on any of the factors and were withheld from further analysis.
The internal consistency of each scale was conputed by coefficient
alpha. Nine items were found to lower the reliability of a scale when
included and had factor loadings less than .39. These items were
withheld from further analysis.

A discriminant analysis was used to construct discriminant
functions for both sets of scales. In the a priori approach, the
discriminant function correctly classified 86% of a randomly selected
subgroup of a sample of depressed subjects and 57% of a group selected
for cross validation. The discriminant function constructed by the
empirical approach correctly classified 77% of the randomly selected
subgroup and 43% of the group selected by cross validation.

A second discriminant analysis was used for both approaches using
scales selected on the basis of the literature review. In the a priori
approach, the discriminant function accurately classified 73% of the
randomly selected groups and 60% of the cross validation groups. In
the empirical approach, the discriminant function accurately classified
73% of the randomly selected groups, and 60% of the cross validation
groups.

The eighteen items that discriminated between the diagnostic
groups were aggregated. A comparison between group means for the
seventy subjects selected at random was significant at the .01 level.
However, there» was no significant difference (p<.05) between group

means for the cross validation groups.

CHAPTER 5
SUMMARY AND CONCLUSIONS

In Chapter 5, four sections are presented. The first section
contains a summary of the research project. The conclusions are drawn
in the second section, and a discussion of the findings is presented in
the third section. The fourth section of the chapter contains impli-

cations for future research.

em

The purpose of the study was to determine whether or not a patient
self rated instrument could differentiate endogenous and non-endogenous
depression. Two major conclusions were drawn from a review of the
literature. Three of the 22 studies failed to support the conclusion
that depression was not a homogenous entity. Symptoms that most
frequently distinguished endogenous from non-endogenous depression were
psychomotor disturbance, autonomous course, self reproach, distinct
quality of mood, and middle or terminal insomnia. The second
conclusion was that although there had been several self rated measures
of depression in the literature, none were designed to differentiate
subtypes of depression.

After the review of the literature was complete, 29 symptom
categories of depression were identified. A research team generated
141 items that were designed to represent these categories. The

resultant instrument was named the Differential Diagnostic Depression

72

73

Scale (ODDS). A sample of 100 individuals with the chief complaint of
depression was obtained through a consortium of inpatient and out-
patient psychiatric centers. The DDDS was administered to each subject
as well as the Dexamethasone Suppression Test (DST) and the Schedule
for Affective Disorders and Schizophrenia (SADS). Individuals were
diagnosed as endogenous or non-endogenous based on a combination of DST
values and Research Diagnostic Criteria (RDC).

In the data analysis, two approaches were used to combine items
into research scales. In the a priori approach, items were grouped
into scales based on similarity in content. The frequency of responses
per diagnostic category was compared by chi-square, and the internal
consistency of each scale was computed. Fifteen items decreased the
internal consistency of their original scale and were not significant
at the .15 level or below. These items were deleted from further
analysis. Ten items decreased the internal consistency and were
significant at the .15 level. These items were placed in a general
category labeled endogenous, or E Scale. Twelve additional items were
deleted to ensure an equal number of items as the empirical approach.

A principal component solution was used in the empirical approach
and 15 factors were extracted. A varimax rotation was used. Eighteen
items failed to load on any of the extracted factors and were withheld
from further analyses. The internal consistency of the factors was
computed and the nine items that lowered the internal consistency and
did not load higher than .39 on their respective factors were deleted.

A discriminant analysis was used to construct a function that
would best separate endogenous and non-endogenous depression for each

approach. Thirty five subjects from each group were randomly selected

74

for the analysis phase. The discriminant function constructed by the a
priori approach correctly classified 87% of the randomly selected
groups, and was significant at the .05 level. The discriminant
function constructed by the empirical approach correctly classified 77%
of the randomly selected groups, and was significant at the .005 level.
Cross validation was attempted by using the discriminant functions to
classify the 30 remaining individuals in the sample. The classifi-
cation accuracy was 57% with the a priori discriminant function, and
44% with the empirical discriminant function.

A second discriminant function was constructed for both approaches
using a reduced number of scales. In the a priori approach, the second
function correctly classified 73% of the original seventy subjects, and
60% of the cross validation groups. In the empirical approach, the
second function correctly classified 73% of the cross validation
groups.

Finally, eighteen items that discriminated between non-endogenous,
probable, and definite endogenous depression were aggregated. There
was a significant difference (p<.01) between the endogenous and
non-endogenous group means for the original seventy subjects. However,
there was no significant difference (p>.05) between the group means for

the cross validation groups.

Conclusions
The conclusions reached from the view of the literature were as
follows:
1. Previous studies on symptom differences between endogenous and
non-endogenous depression had considerable variation in sample
composition, diagnostic criterion, and range of symptoms

included for analysis.

The
1.

5.

75

Nineteen of the 22 studies reviewed supported the conclusion
that depression can best be explained using a binary, or
pluralistic paradigm.

The most frequently cited symptom differences in the litera-
ture between endogenous and non-endogenous depression were
psychomotor disturbance, autonomous course, self reproach,
distinct quality of mood, and middle or terminal insomnia.
conclusions from the study are as follows:

There was a significantly higher percentage of females in the
endogenous sample than in the non-endogenous sample (x2,
p<.05). There were no other significant differences in the
demographics for the two groups.

None of the 141 DDDS items had an endorsement level greater
than .95.

One hundred and fourteen items survived the tests for internal
consistency in the a priori approach. Twenty five scales were
constructed ranging in reliability from .51 to .88.

Fifteen factors emerged from a principal component solution in
the empirical approach. One hundred and fourteen items had
factor loadings above .39 and did not decrease the internal
consistency of the factors. The reliability of the factors
ranged from .31 to .92.

The discriminant function constructed by the a priori approach
correctly classified 87% of a random sample of 70 of the 100
original subjects. 0n cross validation, the correct classifi-

cation shrank to .57.

76

6. The discriminant function constructed by the empirical
approach correctly classified 77% of a random sample of 70 of
the 100 original subjects. On cross validation, the correct

classification shrank to .44.

Discussion

 

The Differential Diagnostic Depression Scale (ODDS) is a self
rated questionnaire. developed as an inexpensive alternative for the
differentiation of subtypes of depression. In the pilot study of the
instrument, two approaches were used to combine items into scales: a
priori classification and factor analysis.

The two discriminant functions constructed from the different
approaches classified the random sample of 70 subjects with impressive
accuracy (87%. and 77%, respectively). However, the accuracy rate
proved to be highly unstable, dropping to 57% for the a priori function
and 44% for the empirical function with the cross validation groups.
The instability can be explained by three possible sources. The first
source is error variance in the scales, or the variables used in the
discriminant analyses to construct the functions. The second potential
source of error was variance in the diagnostic classifications.
In addition, there may have been systematic differences between the two
groups used in the classification and cross validation phases of the
study.

There was considerable variability in the reliability of the
scales used for both approaches. Although attempts were made in both
approaches to maximize the internal consistency of the scales,
considerable inter-scale differences existed in scale reliability. The

alpha coefficients ranged from .51 to .88 with the a priori approach,

77

and .32 to .92 with the empirical approach. The scales in both
approaches with low reliability may have contributed to the instability
of the classification accuracy. One reason for the differences in the
reliability of the scales can probably be attributed to differences in
scale lengths. Reliability of scales increases with longer scales.
Furthermore, some of the content areas that have been associated with
endogenous depression may have been under-represented by items. For.
example, although the review of the literature in Chapter 2 showed that
"distinct quality of mood" was one of the most consistent symptoms of
endogenous depression, it was represented by only one item on the ODDS.
The reason for having only one item was that certain symptom categories
did not readily lend themselves to item generation. The research team
made every reasonable effort to generate items for each category.

A second reason for the variability in the instrument may have
been that the items differed in the reading level required for their
comprehension. No tests were performed on the items to ensure that
items were not above an eighth grade level of reading comprehension,
however, efforts were made to keep the vocabulary of each item as
simple as possible.

Criterion variance may also have contributed to the shrinkage in
classification accuracy between the two sample groups. The endogenous
group was formed by collapsing two diagnostic subtypes: those indivi-
duals with a 'probable' endogenous depression and those with a
'definite' endogenous depression. Some researchers have advocated
eliminating the probable endogenous group from classification studies
because there are false positives in the group (Feinberg, et. al.,

1982; Spitzer, 1978). The inclusion of the probable endogenous group

78

in the study may have had the effect of minimizing the differences
between the two diagnostic groups. Secondly, although the reliability
of the RDC has been established (Spitzer, 1978; Gibbon, 1981; and
Holmes, et. al., 1983) questions remain about the validity of the
diagnostic categories. Carroll has argued that the RDC is not as
accurate in classifying endogenous depression as his clinical
diagnostic method (Carroll, et. al., 1981). Nelson and Charney (1981)
have stated in their review that other diagnostic criteria, such as the
DSM-III, may be more sensitive to endogenous depression.

The phenomena of endogenous depression may be unstable in itself.
Carroll (1982) has argued that the DST may suffer a loss of sensitivity
and specificity if it is not administered in the middle of the episode.
Unfortunately, the lack of research in measuring the episodic nature of
endogenous depression did not permit an accurate classification of the

sample on this variable.

Implications and Recommendations for Further Research
The pilot study of the DDDS was an initial attempt to develop a
self rated instrument that. would differentiate endogenous and
non-endogenous depression. Considering the difficulties discussed
above, further research is strongly recommended as follows:

1. Construction of new discriminant functions with the probable
endogenous group eliminated fron the analysis and classifi-
cation phases. This step would likely increase the separation
of the two groups by removal of the false positives. Further
research is required into the nature of the probable endo-

genous group before they can be included in classification

studies.

2.

79

Refining the scales used in the a priori and empirical
approaches. There are several methods that can be used in
this approach. First, some of the shorter, less reliable
scales should be lengthened and the internal consistency
recalculated. Second, a subset of the scales generated in the
study could be selected for the construction of a new
discriminant function. The selection of the scales should be
based on the following criteria: the scale should be reliable,
have a low correlation with another scale, and discriminate
between the two depressed groups by univariate analysis.

Use the DST and medication response to decrease criterion
variance. Another approach to the diagnostic classification
of endogenous depression is the use the DST and/or medication
response to form the known groups for the discriminant

analysis and classification.

APPENDICES

APPENDIX A

80

Appendix A
Differential Diagnostic Depression Scale (ODDS)
~
All items are to be marked:

OT 8 Definitely Irue of me
T 8 True of me
NT 8 Not True of me
ONT 8 Eefi'n’itely Not lrue of me

(DT) (T) (NT) (ONT)

1. I feel like my heart is racing. () () () ()
2. I don't have any energy. () () () ()
3. I'm so tired I don't care about '
anything anymore. () (l () (l
4. People can read my mind. () () () ()
5. I wake up refreshed. () () () ()
6. I have many different fears about
the present and future. () () () ()
7. I have little interest in being with
friends. () i) (i (l
8. I am careful about my manner of dress. () () () ()
9. Food doesn't taste good to me anymore. () () () ()
10. My chest feels tight. () () () ()
11. I hold myself back from doing what
I want. (1 (l (l (l
12. I watch what I am doing with other
people. () (l (l (l
13. My depression is mild but always there. () () () ()
14. I have problems with constipation. () () () ()
15. I feel my condition is hopeless. () () () ()

16. I can remember “playing sick” to get
out of sonething. () () () ()

17.
18.
19.

20.
21.

22.

23.

24.

25.

26.
27.
28.
29.

30.

31.

32.

33.

34.
35.
36.

81

Appendix A (continued)

People tell me I sigh a lot.
I have problems with diarrhea.

I hear things that other people don't
hear.

Nothing seems to make me feel better.

My depression is so overwhelming, I
find it hard to go on.

My depression came on me quickly.

I have a hard time making a decision
on small things but I do alright on
the important things.

I just can't be courteous to people
who are disagreeable.

My depression started less than a
year ago.

I feel worthless and no good.
I can't stand myself anymore.
I don't like myself.

I have been slowly getting more and
more depressed over the past few weeks.

Other people in my family have had
problems with depression.

I continue to cry over the event that
caused my depression.

Hhen I don't know something I try to
cover it up.

I've made a serious attempt to harm
myself.

I have no hope for myself.
I can't let go and have fun anymore.

I feel nervous and edgy.

(DI)

()
(l

()
()

(l

()

()

(l

(l
()
()
(l

()

(l

(l

(l

(l
(l
()
(i

(T)
()
(l

()
(l

(l

()

(l

(l

()
()
()
i)

()

(l

()

(l

()
(l
()
()

(NT)
()
(l

()
()

(l

()

()

(l

()
()
(l
()

()

()

()

(l

(l
(l
(I
()

(DNT)

()
()

(l
(l

()

()

()

()

(l
(l
(l
()

()

(l

(l

()

()
I)
()
I)

82

Appendix A (continued)
(DT) (T) (NT) (ONT)

37. I can't seem to keep my mind on one

thing. () () (l (l
38. I wake up early in the morning and

can't get back to sleep. () () () ()
39. I can't concentrate. () () () ()
40. I never like to gossip. () () () ()
41. I don't seem to have any control over

my life. () () () (l
42. I tried to commit suicide, but I knew

it wasn't going to work. () (l (l (l
43. I feel so bad I don't feel anything

anymore. () () () ()
44. Everything about me is slowed down. () () () ()
45. I don't take chances anymore. () () () ()
46. Nothing makes me feel better. () () () ()
47. I feel empty and hollow. () () () ()
48. I do not hesitate to go out of my way

to help someone in trouble. () () () ()
49. Things aren't going to get any better

for me. () (l (l (l
50. I'm tired most of the time, but

sometimes I have a burst of energy. () () () ()
51. I seem to cry for no reason at all. () () () ()
52. I feel as if things are not real. () () () ()
53. There have been times in my life when

I had no idea what 1 did. () () () ()
54. I can't remember the last time I had

a good laugh. () (I (l (l
55. I'm afraid of being alone. () () () ()

56. I rarely check the safety of my car no
matter how far I am traveling. () () () ()

57.
58.
59.
60.

61.
62.

63.
64.

65.
66.
67.
68.

69.
70.

71.
72.
73.
74.

75.
76.
77.

78.

83

Appendix A (continued)

I'm more depressed than usual.
I've lost interest in doing my work.
My body feels like its rotting inside.

I've been losing weight recently, even
though I'm not trying to.

I feel like my mind is racing.

I don't see how anybody can stand
being around me.

I have been hOSpitalized for depression.

Sometimes I deliberately hurt someone's
feelings.

I pull away from people.
I have problems with headaches.
I can't count on my memory.

I feel worse in the morning than any
other time in the day.

I have pulled into myself.

Hhen I'm not thinking about what
depresses me, I can think clearly.

I get depressed over nothing.
I am sloppy about my manner of dress.
I feel powerless.

I have the same thoughts over and
over again.

I don't allow myself to get involved.
I feel guilty.

My ability to manage my life changes
a lot.

I have problems going to sleep.

IDT)
I)
I)
I)

I)
I)

I)
I)

I)
I)
I)
I)

I)
I)

I)
I)
I)
I)

I)
I)
I)

I)
I)

(T)
I)
I)
I)

I)
I)

I)
I)

I)
I)
I)
I)

I)
I)

I)
I)
I)
I)

I)
I)
I)

I)
I)

INT)
I)
I)
I)

I)
I)

I)
I)

I)
I)
I)
I)

I)
I)

I)
I)
I)
I)

I)
I)
I)

I)
I)

(our)
I)

I)
I)

I)
I)

I)
I)

I)
I)
I)
I)

I)
I)

I)
I)
I)
I)

I)
I)
I)

I)
I)

79.
80.

81.

82.
83.
84.
85.
86.

87.

88.

89.
90.

91.
92.
93.
94.

95.

96.

97.
98.
99.

84

Appendix A (continued)

No one can help me now.

It is sometimes hard for me to go on
with my work if I am not encouraged.

My depression is so severe I drive
pe0ple away.

I have lost my sense of responsibility.
I get irritated a lot.

I can't do anymore than I'm doing now.
I'm very depressed.

As the days pass, I seem to get more
nervous.

I'm ashamed of the fact that I feel
so tired.

I sometimes try to get even rather
than to forgive and forget.

I'm really tired of crying.

Things I like to do don't seem as
much fun as they used to be.

I have a hard time remembering things.
I have chest pains frequently.
People can't count on me anymore.

I could not survive unless someone
took care of things for me.

Although I feel bad, I can do things
if I have to.

No matter who I'm talking to, I'm
a good listener.

I feel sad.
I've tried to commit suicide.

I've lost my interest in sex.

IDT)

I)

I)

I)
I)
I)
I)
I)

I)

I)

I)
I)

I)
I)
I)
I)

I)

I)

I)
I)
I)
I)

IT)
I)

I)

I)
I)
I)
I)
I)

I)

I)

I)
I)

I)
I)
I)
I)

I)

I)

I)
I)
I)
I)

INT)
I)

I)

I)
I)
I)
I)
I)

I)

I)

I)
I)

I)
I)
I)
I)

I)

I)

I)
I)
I)
I)

(out)
I)

I)

I)
I)
I)
I)
I)

I)

I)

I)
I)

I)
I)
I)
I)

I)

I)

I)
I)
I)
I)

100.
101.

102.
103.

104.
105.
106.

107.

108.
109.
110.
111.
112.

113.

114.
115.
116.

117.

118.

119.

85

Appendix A (continued)

Sometimes I get so nervous I panic.

I am the only one in my family that I

know of who has problems with depression.

I have little desire to eat.

I have felt that people have been making

me do things by controlling my mind.
I do not intensely dislike anyone.

I seem to drive people away.

My depression seemed to come out of
nowhere.

Although people don't believe it, I
know there is something wrong with me.

I am more depressed than anyone I know.
I can't cry even when I feel like it.

I cannot manage my own life.

The smallest things seem to upset me.

There have been occasions when I took
advantage of someone.

Hhen things are going well I feel like
I have more energy.

I wake up a lot during the night.
I get upset easily.

People are there to help me now that
I need them.

I've done things too terrible to
think about.

I don't remember things as well as
other people seem to.

I eat less now than usual.

(PT)

I)

I)
I)

I)
I)
I)

I)

I)
I)
I)
I)
I)

I)

I)
I)
I)

I)

I)

I)
I)

IT)
I)

I)
I)

I)
I)
I)

I)

I)
I)
I)
I)
I)

I)

I)
I)
I)

I)

I)

I)
I)

INT)
I)

I)
I)

I)
I)
I)

I)

I)
I)
I)
I)
I)

I)

I)
I)
I)

I)

I)

I)
I)

(ONT)
I)

I)
I)

I)
I)

I)

I)

I)
I)
I)
I)
I)

I)

I)
I)
I)

I)

I)

I)
I)

120.

121.

122.

123.
124.
125.

126.

127.
128.

129.

130.
131.
132.
133.
134.

135.
136.

137.

138.

86

Appendix A (continued)

I have not deliberately said something
to hurt someone's feelings.

If I weren't so afraid of dying, I'd
commit suicide.

Someone in my family has been
hospitalized for depression.

I feel like my mind is racing.
I feel like my body is speeded up.

I wake up often, but finally fall into
a deep sleep.

I have had difficulties getting close
to people for a long time.

I find it impossible to make a decision.

Before voting, I thoroughly investigate

the qualifications of all the candidates.

I have a lot of hard-to-place aches and
pains.

I want to be alone most of the time.
I can hardly think at all.

I have difficulties breathing.

My stomach seems to be upset a lot.

My depression varies from mild to
severe.

My body is going to pieces.

The urge to tell someone off is not
part of my make-up.

I wake up depressed and feel like
things are going to get worse.

I think of strange things that are too
bad to talk about.

IDT)

I)

I)

I)
I)
I)

I)

I)
I)

I)

I)
I)
I)
I)
I)

I)
I)

I)

I)

I)

IT)

I)

I)

I)
I)
I)

I)

I)
I)

I)

I)
I)
I)
I)
I)

I)
I)

I)

I)

I)

INT)

I)

I)

I)
I)
I)

I)

I)
I)

I)

I)
I)
I)
I)
I)

I)
I)

I)

I)

I)

(ONT)

I)

I)

I)
I)
I)

I)

I)
I)

I)

I)
I)
I)
I)
I)

I)
I)

I)

I)

I)

139.

140.
141.
142.

143.

144.

145.

146.

147.

148.

149.

150.

151.
152.

153.

154.

155.

156.

157.

87

Appendix A (continued)

I can't seem to think about anything
but how bad I feel.

If I could cry, I would feel better.
I need people more now than before.

I'm afraid about what will happen
in the future.

My breathing seems different now.
If I could get into a movie without
paying and be sure I was not seen,
I would probably do it.

I've come to live with the fact that
I'm always tired.

I think about suicide a lot, but I
know I won't do it.

I usually get along with people, but
lately my relationships seem to have
been falling apart.

I know what caused my depression.

I find myself sighing a lot.

I feel tired and jittery when I wake up.
I need to be by myself.

There have been times when I feel like
rebelling against people in authority
even though I knew they were right.

I feel tense.

I feel miserable in the morning, but
have high hopes for the day.

I can't seem to get rid of some of the
thoughts that bother me.

I see things that other people don't
see.

I eat more when I get depressed.

IDT)

I)
I)
I)

I)

I)

I)

I)

I)

I)
I)
I)
I)
I)

I)

I)

I)

I)

I)
I)

IT)

I)
I)
I)

I)

I)

I)

I)

I)

I)
I)
I)
I)
I)

I)

I)

I)

I)

I)
I)

INT)

I)
I)
I)

I)

I)

I)

I)

I)

I)
I)
I)
I)
I)

I)

I)

I)

I)

I)
I)

(DNT)

I)
I)
I)

I)

I)

I)

I)

I)

I)
I)
I)
I)
I)

I)

I)

I)

I)

I)
I)

158.
159.
160.
161.
162.

163.

88

Appendix A (continued)

All the joy is gone out of my life.

My body seems to be working all right.
I admit my mistakes.

I hear strange things when I'm alone.

Most of the time I'm not interested
in sex, but once in a while I am.

My feelings about myself change a lot.

(PT)

I)
I)
I)
I)

I)
I)

(T)
I)
I)
I)
I)

I)
I)

INT)
I)
I)
I)
I)

I)
I)

(ONT)
I)
I)
I)
I)

I)
I)

APPENDIX B

89

Appendix B

Consent Form

Study: Self Reported Symptomatology
*In Major Depressive Illness

Investigator: Gregory Alan Holmes, M.A.

Doctoral Candidate, Counseling Psychology
Robert J. Bielski, M.D.

Department of Psychiatry

I, , have had the above named
study explained to my satisfaction and I freely consent to participate
in the study. I have been informed that my decision to participate in
this study will in no way alter the treatment I will receive for my
depression.

I understand that I have been asked to complete three procedures
during this study. These three procedures are as follows:

1. I will be interviewed by members of the research team.

The interview will last for approximately 1-1/2 hours,
during which time I will be asked questions about my
depression.

2. I will be given the Dexamethasone Suppression Test, a

blood test for plasma cortisol concentration. I will be
given 1 mg of Dexamethasone in a tablet form and be
asked to take the tablet at 11:00 p.m. I understand
that there is minimal risk in taking this medication. I
will then have a small sample of my blood drawn the
following day at 4:00 p.m.

3. I will be given the Differential Diagnostic Depression
Scale (ODDS), a 163 item questionnaire.

I agree to participate in the procedures described above. I
understand that the amount of risk and discomfort involved in this
study is very small; being no greater than that usually involved in

drawing a small blood sample. I understand that the benefits to me

90

Appendix B (continued)

from participating in this study will be a special evaluation of my
depression.

I further understand that I may speak to and/or meet with one of
the study psychiatrists at this time.

I further understand that I may ask questions before signing this
consent form, or any time thereafter, that my participation in this
study is voluntary and that I am free to withdraw at any point without
penalty.

I further understand that the results from these procedures are
confidential and can only be released to others with my written

permission.

Signed:

 

(Subject) (Date)

 

(Hitness) (Date)

xc: Subject
File

APPENDIX C

91

Appendix C

Items Deleted from ODDS
During A Priori Approach

M

 

Item Original Alpha IF:
Number Scale Included/Deleted Significance
140 Crying .68 .82 .75
4 Delusions/Hallucinations .70 .71 .88
157 Eating .70 .77 .34
12 Impulse Inhibition .777 .779 .78
113 Reactivity .50 .66 .74
90 Loss of Interest .762 .763 .24
154 Diurnal Variation .681 .712 .34
41 Hopelessness .813 .855 .63
84 Responsibility .731 .769 .35
155 Obsessiveness .332 .566 .56
83 Sadness .737 .758 .67
125 Insomnia .757 .813 .93
159 Somatic .786 .82 .41
116 Social Hithdrawal .764 .788 .82

141 Social Withdrawal .764 .789 .65

 

APPENDIX D

92

Appendix D

E Scale Items

———_—_———___—_———————-——————-——_————'——-—_-—.

 

Item Original Alpha IF:

Number Scale Included/Deleted Significance
7O Concentration .772 .826 .06
53 Delusions/Hallucinations .704 .704 .15
55 Fear .541 .724 .O4
29 Onset .114 .517 .02
63 History .563 .744 .01

163 Reproach .848 .867 .07
13 Severity .696 .748 .14
121 Suicide .831 .841 .008
147 Social Withdrawal N/A N/A .07

117 Self Reproach N/A N/A .004

 

APPENDIX E

93

Appendix E

Scales Deleted by
‘9" "1°" A_'_'_"_3CI‘_____

   

 

Item Original Scale Significance
23 Decision Making .26
127 Decision Making .58
6 Fear .28
142 Fear .16
97 Sadness .40
158 Sadness .16
151 Social Withdrawal .53
130 Social Withdrawal .68
126 Social Withdrawal .50
105 Social Withdrawal .52

69 Social Hithdrawal .30
65 Social Withdrawal .80

APPENDIX F

94

APPENDIX F. CORRELATION MATRIX FOR A PRIORI SCALES.

 

 

Anxiety Crying Concen D a H Eating Impulse Reactive

 

Anxiety 1.00 .22 .57 .29 :38 .49 .38
Crying .21 1.00 .12 .14 .27 .01 .14
Concen .57 .12 1.00 .18 .26 .43 .51
Schizo .29 .14 .18 1.00 .19 .45 .33
Rating .38 .27 .26 .19 1.00 .33 .34
Impulse .49 .Ol .43 .45 .33 1.00 .56
Reactive .38 .14 .51 .33 .34 .56 1.00
Interest .35 .21 .38 .30 .67 .47 .50
Memory .32 .23 .44 .15 .05 .16 .26
Diurnal .52 .13 .55 .16 .30 .38 .42
Onset .12 -.13 .20 .19 .32 .16 .30
Hopelessness .24 .23 .28 .42 .26 .42 .54
History .12 -.05 .10 .20 -.O7 .17 .05
Psymotor .62 .19 .59 .42 .37 .55 .53
Response .30 .15 .51 .43 .23 .46 .38
Obsess .56 .30 .52 .31 .39 .49 .35
Reproach .61 .29 .45 .43 .31 .58 .55
Severity .44 .21 .55 .33 .31 .46 .69
Insomnia .41 .35 .33 .26 .43 .18 L19
Somatic .41 .42 .41 .44 .40 .37 .42
E Scale .11 .37 .15 .44 .34 .18 .29
Distinct .32 .31 .32 .23 .45 .41 .48
Length .07 .02 -.Ol .18 .31 .23 .IO
Irrate .42 .16 .36 .14 .09 .32 .34
Precip -.06 -.O4 -.16 -.13 .13 -.O6 -.26

Suicide .09 .21 .03 .35 .24 .09 .21

 

95

 

 

 

APPENDIX F. CORRELATION MATRIX FOR A_PRIORI SCALES (CON'T.).
Interest Memory Diurnal Onset Hopelessness History
Anxiety .35 .32 .52 .12 .24 .12
Crying .21 .23 .13 -.03 .23 -.05
Concen .38 .44 .55 .20 .28 .IO
Schizo .29 .15 .16 .19 .42 .20
Eating .67 .OS .30 .32 .26 -.O7
Impulse .47 .18 .38 .16 .42 .17
Reactive .50 .26 .42 .30 .54 .05
Interest 1.00 .11 .35 .24 .46 .05
Memory .11 1.00 .23 -.01 .22 .16
Diurnal .35 .23 1.00 .07 .20 .13
Onset .24 -.01 .07 1.00 .13 -.O9
Hopelessness .46 .22 .20 .13 1.00 .04
History .05 .16 .13 -.09 .04 1.00
Psymotor .56 .26 .51 .14 .36 .06
Response .49 .21 .31 .09 .50 .12
Obsess .40 .34 .43 .21 .32 .16
Reproach .43 .35 .4O .09 .58 .07
Severity .42 .29 .37 .27 .47 .05
Insomnia .45 .04 .32 .11 .21 -.O4
Somatic .58 .17 .40 .19 .45 .05
E Scale .43 .22 .13 .12 .37 .17
Distinct .51 .04 .28 .13 .33 .01
Length .32 -.23 .Ol .34 .14 —.20
Irrate .30 .33 .31 .10 .20 .21
Precip -.04 —.18 -.O7 .14 -.17 -.05
Suicide .26 .16 .08 -.O7 .37 -.03

 

96

APPENDIX F. CORRELATION MATRIX FOR A_PRIORI SCALES (CON'T.).

 

 

Psymotor Response Obsess Reproach Severity Insomnia

 

Anxiety .62 .30 .56 .61 .44 .41
Crying .19 .15 .30 .29 .21 .35
Concen .59 .51 .52 .45 .55 .33
Schizo .42 .43 .31 .43 .33 .26
Eating .37 .23 .39 .31 .31 .43
Impulse .55 .46 .49 .58 .46 .18
Reactive .53 .38 .35 .55 .69 .19
Interest .56 .49 .40 .43 .42 .45
MEmory .26 .21 .34 .35 .29 .04
Diurnal .51 .31 .43 .41 .37 .32
Onset .14 .09 .21 .09 .27 .11
Hopelessness .36 .50 .32 .59 .47 .21
History .06 .12 .16 .07 .05 -.O4
Psymotor 1.00 .44 .46 .60 .48 .37
Response .44 1.00 .40 .49 .56 .33
Obsess .46 .40 1.00 .52 .56 .48
Reproach .59 .49 .52 1.00 .60 .28
Severity .48 .56 -.56 .60 1.00 .36
Insomnia .37 .33 .48 .28 .36 1.00
Somatic .55 .47 .44 .50 .50 .47
E Scale .32 .38 .26 .38 .33 .37
Distinct .43 .23 .40 .31 .38 .36
Length .13 .09 .05 .01 .05 .21
Irrate .35 .32 .37 .42 .42 .18
Precip -.11 -.31 -.O3 -.18 -.28 -.01

Suicide .15 .20 .15 .26 .23 .27

 

97

 

 

 

APPENDIX F. CORRELATION MATRIX FOR A PRIORI SCALES (CON'T.)
Somatic E Scale Distinct Length Irrate Precip Suicide

Anxiety .41 .11 .32 .07 .42 -.06 .09
Crying .42 .37 .31 .02 .16 -.O4 .21
Concen .41 .15 .32 -.Ol .36 -.16 .03
Schizo .44 .44 .23 .18 .14 -.13 .35
Eating .40 .34 .45 .31 .09 .13 .24
Impulse .38 .18 .40 .23 .32 -.O6 .09
Reactive .42 .29 .48 .06 .34 -.26 .21
Interest .58 .44 .51 .32 .30 -.O4 .26
Manory .17 .22 .04 -.23 .33 -.18 .12
Diurnal .40 .13 .28 .Ol .31 -.O7 .08
Onset .19 .02 .13 .34 .10 -.14 -.O7
Hopelessness .45 .37 .33 .14 .20 -.17 .37
History .05 .17 .Ol -.20 .21 -.05 -.O3
Psymotor .55 .32 .43 .13 .38 -.11 .15
Response .47 .38 .23 .09 .32 -.31 .20
Obsess .44 .26 .40 .05 .37 -.O3 .15
Reproach .50 .37 .31 .Ol .42 -.18 .26
Severity .50 .33 .38 .05 .42 -.28 .23
Insomnia .47 .37 .36 .21 .18 .Ol .27
Somatic 1.00 .47 .40 .25 .29 -.15 .26
E Scale .47 1.00 .32 .08 .14 -.O7 .59
Distinct .40 .32 1.00 .33 .19 .02 .16
Length .25 .08 .33 1.00 -.03 .13 -.01
Irrate .29 .14 .19 -.03 1.00 -.31 .07
Precip —.15 8.07 .02 .13 -.31 1.00 .07
Suicide .26 .59 .16 -.01 .07 .07 1.00

 

 

APPENDIX G

98

Appendix G

Factor Items Deleted

 

 

 

 

 

 

Item

Number
79 Suicide Ideation/Attempts .861 .857
153 Agitation .803 .802
125 Agitation .822 .801
114 Dependency .819 .817
77 Interpersonal Dissatisfaction .832 .831
25 Memory Impairment .601 .425
71 Memory Impairment .616 .425
151 Loss of Energy .722 .70
5 History .729 .693

 

 

 

 

 

APPENDIX H

99

APPENDIX E. CORRELATION MATRIX FOR.EMPIRICAL FACTORS.

 

 

 

Factor 1 Factor 2 Factor 3 Factor 4 Factor 5 Factor 6
Factor 1 1.00 .67 .29 .52 .34 .53
Factor 2 .67 1.00 .22 .47 .26 .55
Factor 3 .29 .22 1.00 .17 .29 .26
Factor 4 .52 .47 .16 1.00 .42 .46
Factor 5 .34 .26 .29 .42 1.00 .25-
ictor 6 .63 .55 .26 .46 .25 1.00
Factor 7 .58 .56 .13 .32 .11 .55
Factor 8 .37 .38 .11 .21 -.O4 .25
Factor 9 .25 .31 .29 .32 .25 .23
Factot 10 .41 .46 .27 .46 £38 .51
Factor 11 .36 .37 .39 .25 .14 .31
Factor 12 .34 .12 .10 .16 .26 .07
Factor 13 .48 .42 .15 .16 .07 .23
Factor 14 .14 .16 .06 .06 -.05 .11
Factor 15 .31 .16 .15 .28 .15 .34

 

100

APPENDIX B. CORRELATION MATRIX FOR EMPIRICAL FACTORS (cont.)

 

 

Factor 7 Factor 8 Factor 9 Factor 10 Factor 11 Factor 12

 

Factor 1 .58 .36 .25 .41 .36 .34
Factor 2 .53 .37 .31 .46 .37 .12
Factor 3 .13 .11 .29 .27 .39 .33
Factor 4 .32 .21 .32 .46 .25 .10
Factor 5 .11 -.O4 .25 .38 .14 .16
Factor 6 .55 .25 .23 .51 .31 .26
Factor 7 1.00 .29 .14 .38 .31 .07
Factor 8 .29 1.00 .27 .16 .13 .02
Factor 9 .14 .27 1.00 .37 .22 .05
Factor 10 .38 .16 .37 1.00 .32 .34
Factor 11 .31 .13 .22 .32 1.00 .23
Factor 12 .07 .02 .04 .34 .23 1.00
Factor 13 .36 .25 .05 .29 .27 .19
Factor 14 .24 . 23 .06 .04 . 25 .10

Factor 15 .12 .03 .29 .21 .29 .17

101

APPENDIX B. CORRELATION MATRIX FOR EMPIRICAL FACTORS (cont.)

 

 

Factor 13 Factor 14 Factor 15

 

Factor 1 .48 .14 .14
Factor 2 .42 .16 .16
Factor 3 .15 .06 .15
Factor 4 .16 .06 .28
Factor 5 .07 .05 .15
Factor 6 .23 .11 .34
Factor 7 .36 .24 .12
Factor 8 .25 .23 .03
Factor 9 .05 .06 .29
Factor 10 .29 .04 .21
Factor 11 .27 .25 .29
Factor 12 .19 .10 .17
Factor 13 1.00 -.01 .14
Factor 14 -.01 1.00 .03

Factor 15 .14 .03 1.00

BIBLIOGRAPHY

BIBLIOGRAPHY

Akiskal, H.S., Bitar, A.H., Puzantin, V.R., et. al. The nosological
status of neurotic depression. Archives of General Psychiatry,
1978, 35, 756-766.

Andreasen, N.C., and Grove, W.M. The classification of depression:
traditional versus mathematical approaches. American Journal of
Psychiatry, 1982, 139, 45-52.

 

 

Bielski, R.J., and Friedel, R.O. Prediction of tricyclic antidepres-
sant response. Archives of General Psychiatry, 1976, 33,
1479-1489.

Carroll, B.J. The dexamethasone supression test for melancholia.
British Journal of Psychiatry, 1982, 140, 292-304.

 

Carroll, B.J., Feinberg, M., and Greden, J.R., et. al. A specific
laboratory test for the diagnosis of melancholia. Archives of
General Psychiatry, 1981, 38, 15-22.

 

 

Crowne, D.P., and Marlowe, D. The approval motive: studies in
evaluative dependence. New York: John Wiley, 1964.

Daly, R.J., and Cochran, D.M. Affective disorder taxonomies in
middle-aged females. British Journal of Psychiatry, 1970, 117,
257-266.

 

Endicott, J., and spitzer, R. A diagnostic interview: the schedule for
affective disorders and schizophrenia. Archives of General
Psychiatry, 1978, 35, 837-844.

 

 

Eysenck, H.J. The classification of depressive illness. British
Journal of Psychiatry, 1970, 117, 241-250.

 

Feinberg, M., and Carroll, B.J. Separation of unipolar endogenous
depression from non-endogenous depression. British Journal of
Psychiatry, 1982, 384-391.

 

Feinberg, M., Carroll, B.J., Smouse, P.E., and Rawson, S.G. The
Carroll rating scale for depression: III. Comparison with other
instruments. British Journal of Psychiatry, 1981, 138, 205-209.

 

Freud, S. Mourning and melancholia. In Collected Papers. London:
Hogarth Press, 1956.

 

102

103

Garney, M.P., Roth, M., and Garside, R.F. Diagnosis of depressive
syndromes and the prediction of ECT response. British Journal of
Psychiatry, 1965, 111, 659-674.

 

 

Garside, R.F., Kay, D.K., Hilson, I.C., Deraton, I.D., and Roth, M.
Depressive syndromes and the classification of patients.
Psychological Medicine, 1971, 1, 333-338.

 

Gibbon, M., McDonald-Scott, P., and Endicott, JJ. Mastering the art of
research interviewing. Archives of General Psychiatry, 1981, 38,
1259-1262.

 

Gillespie, R.D. The clinical differentiation of types of depression.
prs Hospital Reports, 1929, 79, 306-344.

 

Gurney, C., Roth, M., Kerr, T.A., and Schapira, K. The bearing of
treatment on the classification of affective disorders. British
Journal of Psychiatry, 1970, 117, 251-255.

 

Hamilton M., and Whie, J.M. Clinical syndrome in depression states.
Journal of Mental Science, 1959, 105, 985-997.

Holmes, G., Smithson, S., Shafer, C., and Bielski, 8. Teaching
students to diagnose depression: a pilot study of research
diagnostic criteria. Professional Psychology: Theory and
Practice, 1983, in press.

 

, 1, Hughes, J.R., O'Hara, M.W., and Rehm, C.P. Measurement of depression

in clinical trials: an overview. Journal of Clinical Psychiatry,
1982, 43, 85-88.

 

Kay, D.K., Garside, R.F., Beamish, P., and Roy, J.R. endogenous and
neurotic syndromes of depression: a factor analytic study of 104
cases. British Journal of Psychiatry, 1969, 115, 377-388.

 

Keller, M.B., Klerman, G.L., Lavori, P.W., Fawcett, J.A., Coryell, W.,
and Endicott, J. Treatment received by depressed patients.
Journal of the American Medical Association, 1982, 248, 2848-2855.

 

Kendell, R.F. The classification of depression: a review of
contemporary confusion. British Journal of psychiatry, 1976, 129,
15-28.

 

Kendell, R.F. The classification of depressive illness. London,
Hogarth Press, 1968.

 

Kendell, R.E., and Gourlay, J. The clinical distinction between
psychotic and neurotic depressions. British Journal of
Psychiatry, 1970, 117, 257-266.

 

Kiloh, L.G., Andrews, G., Wilson, M., and Biandin, G.N. The
relationship of the syndromes called endogenous and neurotic
depression. British Journal of Psychiatry, 1972, 121, 183-196.

104

Kiloh, L.G., and Garside, R.F. The independence of neurotic depression
and endogenous depression. British Journal of Psychiatry, 1963,
109, 415-463.

Klein, D.F. Endogenomorphic depression. Archives of General
Psychiatry, 1974, 31, 447-454.

 

 

Klenman, G.L. Clinical research 'hi depression. Archives of General
Psychiatry, 1972, 24, 305-319.

 

 

Klenman, G.L., Endicott, J., Spitzer, R., Hirschfield, R.A. Neurotic
depressions: a systematic analysis of multiple criteria and
meanings. American Journal of Psychiatry, 1979, 136, 57-61.

 

Kraepelin, E. Manic Depressive Insanity and Paranoia. Edinburgh: E
and S Livingston LTD, 1921.

Lewinsohn, P.M., Zeiss, A.H., Zeiss, R.A., and Haller, R. Endogeneity
and reactivity as orthogonal dimensions in depression. Journal of
Nervous and Mental Diseases, 1977, 164, 327-332.

 

 

Lewis, A.J. States of depression: their clinical and aetiological
differentiation. British Medical Journal, 1938, 2, 875-878.

 

MacFaden, H.W. The classification of depressive disorders. Journal of
Clinical Psychology, 1963, 109, 415-463.

 

 

Matussek, P., and Luks, 0. Themes of endogenous and non-endogenous
depressions. Psychiatry Research, 1981, S, 235-242.

 

Matussek, P., Soldner, M., and Nagel, 0. Identification of the
endogenous depressive syndrome based. on the symptoms and
characteristics of the course. British Journal of Psychiatry,
1981, 138, 361-372.

 

McConaghy, N., Joffe, A.D., and Murphy, 8. The independence of
neurotic and endogenous depression. British Journal of
Psychiatry, 1967, 113, 749-784.

 

 

Nelson, J.C., and Charney, 0.8. Primary affective disorder criteria
and the endogenous-reactive distinction. Archives of General
Psychiatry, 1980, 37, 787-793.

 

 

Nelson, J.C., and Charney, 0.5. The symptoms of major depressive
illness. American Journal of Psychiatry,.l978, 35, 773-782.

Overall, J.E., Hollister, L.E., Johnson, M., et. Ial. Nosology of
depression and differential response to drugs. Journal of the
American Medical Association, 1966, 195, 946-948.

 

Prusoff, B.A., Klerman, G.L., and Paykel, E.S. Concordance between
clinical assessments and patients' self report in depression.
Archives of General Psychiatry, 1972, 26, 546-552.

 

105

Raskin, A., Schulterbrandt, J.G., Reatig, N., and McKeon, J.J.
Differential response to chlorpromazine, imiprimine, and placebo.
Archives of General Psychiatry, 1970, 23, 164-173.

'2 .1 Rhem, L.P. Assessment of Depression. In M. Hersen and A. Bellack
(Eds.), Behavioral assessment: a practical handbook. New York:
PergamonFPress, 1976.

Roth, M. Depressive states and their borderlands: classification,
diagnosis, and treatment. Comprehensive Psychiatry, 1960, 1,
135-155.

 

Rosenthal, S.H., and Gudeman, J.E. The endogenous depressive pattern.
Archives of General Psychiatry, 1967, 16, 241-249.

 

Rosenthal, S.H., and Klerman, G.L. Content and consistency in the
endogenous depressive patterns. British Journal of Psychiatry,
1966, 112, 471-484.

 

Spitzer, R., Endicott, J. and Robins, E. Research diagnostic criteria:
rationale and reliability. Archives of General Psychiatry, 1978,
35, 773-782.

 

Weckowicz, T.E., Cropley, A.J., and Muir, W. An attempt to replicate
the results of a factor analytic study in depressed patients.
Journal of Clinical Psychology, 1971,27, 30-31.

I

J. «Ilfllnl
. Win”...