Thought, Language, and Communication Disorders II.
Diagnostic Significance
Nancy C. Andreasen,
MD
\s=b\ This investigation evaluates the frequency of various subtypes of thought, language, and communication disorders in 113 patients with diagnoses of mania, depression, and schizophrenia. It indicates that some types of thought disorder consid-
ered important occur so infrequently as to be of little diagnostic value, such as neologisms or blocking. The traditional concept
of thought disorder, which emphasizes associative loosening, is also of little value, since associative loosening occurs frequently in mania as well as in schizophrenia. This investigation demonstrates that associative loosening can no longer be considered pathognomonic of schizophrenia. On the other hand, an approach that defines various subtypes of thought disorder and uses a concept of negative-vs-positive thought disorder does often permit a distinction between mania and schizophrenia. It is recommended that the practice of referring globally to "thought disorder," as if it were homogeneous, be avoided in the future and instead that the specific subtypes occurring in particular patients be noted in both clinical practice and research. (Arch Gen Psychiatry 36:1325-1330, 1979)
Bleulerian approach to schizophrenia, which makes characteristic "associative loosening" the pathognomonic has been the most influential in American for psychiatry many years.1 Recently, however, both the Bleulerian approach and the specificity of thought disorder have been questioned. Because Bleulerian criteria are difficult to operationalize, other approaches that give prominence to nonaffective delusions and hallucinations rather than to thought disorder have been developed.24 A number of investigators have also observed that thought disorder may occur in other diagnostic groups, such as manic patients, and that abnormalities in thought, language, and communication may occur in people who do not meet criteria for any psychiatric illness, thus implying that thought disorder is a continuous rather than discrete phenomenon and that it shades gradually into normal¬ ity.51" Further, some investigators have also observed that
Thesymptom, a
Accepted for publication Sept 20, 1978. From the Department of Psychiatry, University of Iowa College of Medicine, Iowa City. Reprint requests to Department of Psychiatry, University of Iowa College of Medicine, 500 Newton Rd, Iowa City, IA 52242 (Dr Andreasen).
schizophrenic patients display thought disorder, thereby questioning its diagnostic specificity still more.11 Progress in thought disorder research has been slowed by the diversity of competing conceptualizations and defi¬ nitions, which have made clinical assessments of thought disorder notoriously unreliable. A previous article has presented definitions of the terms commonly used to describe abnormalities in thought, language, and commu¬ nication.12 These definitions have good to excellent interrater reliability. This set of definitions takes a relatively broad approach to the concept of "thought disorder." That is, it conceptualizes "thought disorder" as consisting of a variety of subtypes. It includes a wide range of linguistic and cognitive behaviors, some of which are sometimes considered specific to schizophrenia, such as derailment or neologisms, and others of which might be expected to occur in a broad range of diagnostic groups, such as circumstan¬ tiality or loss of goal. This set of definitions is clinical, empirical, and atheoretical. It permits an investigation of the types of linguistic and cognitive abnormalities that occur in psychiatric patients and the frequency with which they occur. This set of definitions can therefore be used to examine a series of questions concerning the relationship between "thought disorder" and diagnosis. What is the frequency of the various subtypes among various diagnos¬ tic groups? Is any specific subtype pathognomonic of a particular diagnostic group? Is there a particular combina¬ tion or clustering of subtypes that suggests a particular diagnosis? How valid is "thought disorder" as a pathogno¬ monic diagnostic criterion? not all
METHODS
patients was Thought, Language,
based on the Scale for the Assess¬ and Communication (TLC), devel¬ oped by and available from the author. This instrument contains definitions of 18 subtypes of thought disorder often described or observed in psychiatric patients. These definitions and their interrater reliability have been reported in a previous article.12 The scale contains both the definitions and instructions for rating severity of a 0-to-3 or 0-to-4 scale (depending on the item). The rating scale for severity of each subtype has been made as specific and operational as possible; usually, judgments as to severity depend on the frequency with which a particular phenomenon has been observed. For example, the severity scale for poverty of speech is as follows:
Evaluation of
ment of
Downloaded From: http://archpsyc.jamanetwork.com/ by a Yale University User on 05/15/2015
0 No
poverty of speech. A substantial and appropriate number of
replies
questions include
additional information. 1 Slight poverty of speech. Occasional replies do not include elaborated information even though this is appropriate. 2 Moderate poverty of speech. Some replies do not include appro¬ priately elaborated information, and many replies are monosyl¬ labic or very brief ("Yes." "No." "Maybe." "Don't know." "Last to
week."). 3 Severe poverty of speech. Answers are rarely more than a few words in length and never include additional information. Some questions left unanswered. 4 Extreme poverty of speech. Patient is essentially mute.
Ratings for severity, since they tend to be quantitative, are based standard 45-minute interview, and appropriate corrections are made if the actual evaluation interview is longer or shorter. In addition to the 18 subtypes of thought disorder, the scale also contains instructions for making a global rating. Weighted for the global rating is .89. This scale was used to evaluate the subtypes and severity of thought disorder occurring in a total of 113 patients, including 32 manies, 36 depressives, and 45 schizophrenics. To obtain a large sample and to maximize variance within the sample studied, these patients were drawn from three different hospitals that tended to have somewhat different populations. Twenty-four manies, 31 depressives, and 21 schizophrenics were evaluated at Iowa Psychiatric Hospital in Iowa City, an acute treatment facility with a two-week average length of stay, that treats both private and nonprivate patients. Four manies, four depressives, and 23 schizo¬ phrenics were drawn from Mt Pleasant State Hospital, an acute treatment facility for state patients with a mean length of stay of two months. A small number of patients at the Iowa City Veterans Administration (VA) Hospital were also evaluated: five manies, no depressives, and one schizophrenic. Patients from these facilities were included in the study if they met the Research Diagnostic Criteria for manic disorder, major depressive disorder, or schizophrenia (subacute, subchronic, and chronic types).' Although all the schizophrenic patients had had a period of illness greater than six months (including the prodromal period), none were chronic in the sense of having had long-term hospitalizations. The median number of hospitalizations was 2.5 for the schizophrenics, 3 for the manies, and 1 for the depressives. All of the schizophrenics, 84% of the manies, and 62% of the depressives had had prior hospitalizations. Patients were drawn from the Mt Pleasant and Iowa City populations on the basis of consecutive admissions that met the diagnostic inclusion criteria. We began using patients from the VA Hospital toward the end of the study, and there we also used consecutive admissions. We evaluated most patients during the first week after admis¬ sion. Nearly all were receiving medication at the time of evalua¬ tion, since medications are usually prescribed at these facilities as early as possible after admission to reduce length of stay. Although nearly all were receiving medication, all were still severely symptomatic, so that all manifested a full and typical manic, schizophrenic, or depressive syndrome at the time of the evaluation. In a few instances (limited exclusively to manic patients), patients who were admitted consecutively and met inclusion diagnostic criteria had responded to medication so quickly that they no longer manifested the full manic syndrome, and these patients were therefore not included in the study. The three diagnostic subgroups did not differ significantly from one another in age (F 0.158) or educational level (F 0.096). They did differ slightly on sex ratio using stringent criteria on a
=
=
< .135). The manies included 14 men and 18 women age of 33.0 and a mean education of 13.5 years. The depressives included 14 men and 22 women with a mean age of 39.7 and a mean education of 12.7 years. Twenty-seven schizophrenics were male and 18 female; this group had a mean age of 29.0 years and a mean education of 12.1 years. An attempt was made to give all subjects the Shipley Hartford test to assess vocabulary and
(2
=
with
4.000,
a mean
intelligence, but many of the schizophrenics and manies were too ill to cooperate fully with this task, often only completing a few items and then becoming restless. Consequently, these data were considered invalid and were not analyzed. The patients were interviewed and evaluated with the TLC Scale in two different ways. The first set of patients, a total of 44, were evaluated by means of tape-recorded interviews that had been collected in an earlier project. These interviews were approx¬ imately 20 minutes in length, not formally structured, and covered such areas as recent experiences, current events, interests, and
interpersonal relationships. Psychiatric symptomatology
was
not
discussed. These interviews were evaluated with the TLC Scale both by a blind rater (N.C.A.) and by the person who originally conducted the interview. Ratings tended to be quite close, but the data from the blind rater only are reported in this study. The second set of patients (19 manies, 20 depressives, and 30 schizo¬ phrenics) were interviewed for 45 minutes, using a standard structured interviewing procedure. This procedure began by asking the patient to talk about himself for five to ten minutes without interruption (except for prompting to keep him going as needed), covering such areas as his work, home, interests and hobbies, and personality traits. Thereafter, a standard series of follow-up questions was used, ranging from abstract to concrete,
open-ended to closed, and personal to impersonal. Psychiatric symptomatology was not discussed. Ratings were done at the conclusion of the interview by two raters, one of whom had been the interviewer and the other acting as observer. Although the interview was tape-recorded, the tape was not reviewed prior to making ratings. The patients were evaluated initially by a third person, who determined diagnosis, described the study to them, and obtained informed consent. Consequently, both raters were kept as blind as possible concerning diagnosis in the second portion of the study. Data from the two sets of subjects did not differ statistically significantly, and consequently the data from the two sets are pooled in this report. Data were analyzed with use of packaged programs from the Statistical Package for the Social Sciences or Statistical Analysis System. RESULTS Tables 1 and 2 contain simple descriptive data concern¬ ing the frequency and mean ratings of the various subtypes of thought disorder. The frequency counts in Table 1 are based on a rating of 1 or higher on the TLC Scale. A rating of 1 indicates a mild disorder, often as little as a single instance of the item being rated. In spite of this low threshold for positive ratings of these items, many of the subtypes occurred quite infrequently, including some that are often considered very important indicators of formal thought disorder, such as blocking, clanging, neologisms, and incoherence. Many of the subtypes occurred with nearly equal frequency in manies and schizo¬ phrenics, including tangentiality, derailment, incoherence, illogicality, and loss of goal. Nearly all manies and schizo¬ phrenics also had a global rating of 1 or greater. Some subtypes were commoner in mania, such as pressure of speech, clanging, distractible speech, or circumstantiality. Other subtypes occurred more frequently in schizophrenia, such as poverty of speech and poverty of content of speech. Although 53% of the depressed patients had a global rating of 1 or greater, this rating reflected subtypes that were either mild or consistent with other depressive symptoms, such as poverty of speech, poverty of content of speech, tangentiality, circumstantiality, or loss of goal. Table 2 indicates more clearly that most subtypes of thought disorder occurred in a mild form or did not occur at all in depressed patients. The transformed mean global
Downloaded From: http://archpsyc.jamanetwork.com/ by a Yale University User on 05/15/2015
Table
1.—Frequency of Types of Thought in Psychiatric Patients* No.
Derailment Incoherence
Illogicality Clanging Neologisms Word
Subsets!
Schizophrenics ( = 45) 13(29)
6(19) 23(72) 10(31) 11(34)
6(17) 2(6) 0(0) 9(25)
18(56) 5(16) 8(25) 3(9) 1(3)
5(14) 0(0) 0(0) 0(0) 0(0)
18(40) 12(27) 1(2) 16(36) 25(56) 7(16) 12(27) 0(0) 1(2)
1(3) 8(25) 14(44) 11(34) 1(3) 1(3) 2(6) 7(22) 28(88)
0(0) 11(31) 6(17) 2(6) 0(0) 2(6) 1(3) 4(11) 19(53)
32) 2(6) =
Circumstantiality Loss of goal Perseveration Echolalia
Blocking Stilted speech Self-reference Global rating on
ratings
of >1
on
_
Perseveration Global rating
Mean
Poverty of speech Poverty of content of speech Pressure of speech Distractible speech
Tangentiality Derailment Incoherence
Illogicality Clanging Neologisms Word
approxima¬
tions
Circumstantiality Loss of goal Perseveration Echolalia
Blocking Stilted Speech Self-reference Global rating
32)
phrenics ( = 45)
Mean
SD
3.9
8.0
4.7
8.4
2.2 14.1 3.8 4.4 14.4
4.9 1.6 6.1 6.7 15.2 7.8 7.6 4.2 1.8
2.2 0.5 0.0 2.8 1.7 0.0 0.0 0.0 0.0
5.4
7.6 3.6 0.2 5.5 12.2 3.3 4.2 0.0 0.2
10.9 6.5 1.5 8.4 13.3 9.0 8.7 0.0 1.5
1.8 9.2 10.8 8.0 5.3
0.0
0.0
4.2 3.1 0.1 0.0
0.0 0.4 7.1 1.3 0.7 0.6 0.2
0.0 2.1
1.6 18.0
4.1 1.3 0.3
0.3 4.6 8.4 5.0 0.9 0.3 1.3 3.1
20.0
2.3 0.0 5.1 4.5 0.0 0.0 0.0 0.0
4.9
0.2
6.9 7.5 2.3 0.0 2.3 1.6
6.4 1.1
1.1 7.0
3.2 7.4
0.5
*Scores were transformed to range from 0 to 40 by using
constant of 10.
a
SD
9.2 6.4
3.3 3.3 1.5 4.2 10.6
multiplication
7 in the depressed patients, as compared with 20 for the manies and 18 for the schizophrenics. Depressed patients tended to have poverty of speech and to be more circumstantial. Manic patients, on the other hand, tended to have their highest mean ratings on pressure of speech, derailment, tangentiality, illogicality, and loss of goal; they
rating is
D/M.S
.006 .042
S/D.M
4.790
.010
18.163
0.000
D,S/S,M D/M.S
6.435; for setwise
=
=
=
=
highest,
last.
relatively high mean ratings on distractible speech, circumstantiality, and perseveration. Further, the schizophrenic patients had considerably higher mean ratings on poverty of speech and poverty of content of speech. Table 3 indicates the results of an analysis of variance of some of the subtypes of thought disorder that occurred more frequently. Pressure of speech, derailment, distracti¬ ble speech, and the global rating have extremely high F ratios. The Scheffé procedure indicates that part of the variance for pressure of speech and distractible speech derives from differences between manies and schizophren¬ ics, while they tend to resemble one another and to differ from the depressives in terms of derailment and global rating. Poverty of content of speech and circumstantiality are other variables on which manies and schizophrenics tend to differ significantly, and there is a suggestion that poverty of speech might also be useful in differentiating these two groups. The three groups do not differ signifi¬ cantly on any other subtypes if a strict Bonferroni approach is used. These trends suggest that manies and schizophrenics tend to resemble one another in terms of illogicality, perseveration, and loss of goal. These data were also subjected to multivariate analysis of variance (MANOVA) using ten variables: poverty of speech, poverty of content of speech, pressure of speech, distractible speech, tangentiality, derailment, incoherence, illogicality, loss of goal, and perseveration. The generalized F value for this MANOVA was 8.495, with < .01. Two follow-up tests, which had been planned a priori, were done also had
6(13) 41(91)
Mean
=
disorder is listed first and the
2(4) 1(2)
Depressives ( = 36)
D.M/S D.S/M D.S/M
.012_D,M/M,S
=
2.5
3.1
.003 .000 .000 .213 .000 .079
=
0.6
SD
6.089 30.163 13.341 1.566 11.596 2.600
separately: for setwise a .025, critical .05, critical F 5.667. f M indicates manies ( 32); D, depressives ( 36); S, schizophren¬ ics ( 45). The virgule separates groups significantly different from one another; the group with the lowest score for the subtype of thought
Schizo¬
(
.044
*P is calculated for each F ratio
F
0(0) 2(4) 20(44) 11(24) 2(4)
Table 2.—Transformed Mean Ratings for Types of Thought Disorder*
=
3.217
Illogicality _4.602 5.309 Circumstantiality Loss of goal 3.270
TLC Scale.
Manies
Poverty of speech Poverty of content of speech Pressure of speech Distractible speech Tangentiality Derailment. Incoherence
approxima¬
tions
*Based
of Patients
Depressives ( = 36) 8(22)
Manies
( Poverty of speech Poverty of content of speech Pressure of speech Distractible speech Tangentiality
(%)
Table 3—Analysis of Variance and Scheffé Procedure for Differences Between Subsets on Types of Thought Disorder
Disorder
the MANOVA. These are summarized in Table 5. These follow-up tests can be used to compare two different ways of conceptualizing "thought disorder." The first test evalu¬ ates the concept of loose associations by giving an equal positive weighting to each of the five subtypes considered to represent various forms of associative loosening (tan¬ gentiality, derailment, incoherence, illogicality, and clang¬ ing). This follow-up test yields an extremely low F value and does not disclose any significant differences between the three groups in terms of associative loosening. An alternative approach was to conceptualize "thought disoron
Downloaded From: http://archpsyc.jamanetwork.com/ by a Yale University User on 05/15/2015
Table 4.—Variance in Manic and Schizophrenic Patients Accounted for by Discriminant Variables Entered Stepwise
(N
-
77)*
Unstandardized
Coefficients! Pressure of speech Distractible speech Derailment
.61578 .95629 .28826
% of Variance 40.251 17.815 16.457
Cumulative % 58.066 74.523
Poverty of content 12.300 86.823 of speech _.37874 6.808 Poverty of speech_.37843 _93.631 .22657 4.984 Loss of goal 98.615t *These other variables were insufficiently discriminatory and were therefore not entered into the analysis: perseveration, tangentiality, inco¬ herence, and illogicality. tConstant .88329. JDoes not add to 100% due to rounding error. =
Table
5.—Follow-up Tests of Multivariate Analysis of Variance (N 113) =
Loose associations* +1
Tangentiality Derailment Incoherence
+1
Illogicality Clanging Positive and negative formal thought disordert Poverty of speech Poverty of content of speech Pressure of speech
+ 1
Tangentiality
+1
Derailment Incoherence
+1
Illogicality •F
=
fF
=
0.4248, 9.918,
=
Diagnostic Significance
Nancy C. Andreasen,
MD
\s=b\ This investigation evaluates the frequency of various subtypes of thought, language, and communication disorders in 113 patients with diagnoses of mania, depression, and schizophrenia. It indicates that some types of thought disorder consid-
ered important occur so infrequently as to be of little diagnostic value, such as neologisms or blocking. The traditional concept
of thought disorder, which emphasizes associative loosening, is also of little value, since associative loosening occurs frequently in mania as well as in schizophrenia. This investigation demonstrates that associative loosening can no longer be considered pathognomonic of schizophrenia. On the other hand, an approach that defines various subtypes of thought disorder and uses a concept of negative-vs-positive thought disorder does often permit a distinction between mania and schizophrenia. It is recommended that the practice of referring globally to "thought disorder," as if it were homogeneous, be avoided in the future and instead that the specific subtypes occurring in particular patients be noted in both clinical practice and research. (Arch Gen Psychiatry 36:1325-1330, 1979)
Bleulerian approach to schizophrenia, which makes characteristic "associative loosening" the pathognomonic has been the most influential in American for psychiatry many years.1 Recently, however, both the Bleulerian approach and the specificity of thought disorder have been questioned. Because Bleulerian criteria are difficult to operationalize, other approaches that give prominence to nonaffective delusions and hallucinations rather than to thought disorder have been developed.24 A number of investigators have also observed that thought disorder may occur in other diagnostic groups, such as manic patients, and that abnormalities in thought, language, and communication may occur in people who do not meet criteria for any psychiatric illness, thus implying that thought disorder is a continuous rather than discrete phenomenon and that it shades gradually into normal¬ ity.51" Further, some investigators have also observed that
Thesymptom, a
Accepted for publication Sept 20, 1978. From the Department of Psychiatry, University of Iowa College of Medicine, Iowa City. Reprint requests to Department of Psychiatry, University of Iowa College of Medicine, 500 Newton Rd, Iowa City, IA 52242 (Dr Andreasen).
schizophrenic patients display thought disorder, thereby questioning its diagnostic specificity still more.11 Progress in thought disorder research has been slowed by the diversity of competing conceptualizations and defi¬ nitions, which have made clinical assessments of thought disorder notoriously unreliable. A previous article has presented definitions of the terms commonly used to describe abnormalities in thought, language, and commu¬ nication.12 These definitions have good to excellent interrater reliability. This set of definitions takes a relatively broad approach to the concept of "thought disorder." That is, it conceptualizes "thought disorder" as consisting of a variety of subtypes. It includes a wide range of linguistic and cognitive behaviors, some of which are sometimes considered specific to schizophrenia, such as derailment or neologisms, and others of which might be expected to occur in a broad range of diagnostic groups, such as circumstan¬ tiality or loss of goal. This set of definitions is clinical, empirical, and atheoretical. It permits an investigation of the types of linguistic and cognitive abnormalities that occur in psychiatric patients and the frequency with which they occur. This set of definitions can therefore be used to examine a series of questions concerning the relationship between "thought disorder" and diagnosis. What is the frequency of the various subtypes among various diagnos¬ tic groups? Is any specific subtype pathognomonic of a particular diagnostic group? Is there a particular combina¬ tion or clustering of subtypes that suggests a particular diagnosis? How valid is "thought disorder" as a pathogno¬ monic diagnostic criterion? not all
METHODS
patients was Thought, Language,
based on the Scale for the Assess¬ and Communication (TLC), devel¬ oped by and available from the author. This instrument contains definitions of 18 subtypes of thought disorder often described or observed in psychiatric patients. These definitions and their interrater reliability have been reported in a previous article.12 The scale contains both the definitions and instructions for rating severity of a 0-to-3 or 0-to-4 scale (depending on the item). The rating scale for severity of each subtype has been made as specific and operational as possible; usually, judgments as to severity depend on the frequency with which a particular phenomenon has been observed. For example, the severity scale for poverty of speech is as follows:
Evaluation of
ment of
Downloaded From: http://archpsyc.jamanetwork.com/ by a Yale University User on 05/15/2015
0 No
poverty of speech. A substantial and appropriate number of
replies
questions include
additional information. 1 Slight poverty of speech. Occasional replies do not include elaborated information even though this is appropriate. 2 Moderate poverty of speech. Some replies do not include appro¬ priately elaborated information, and many replies are monosyl¬ labic or very brief ("Yes." "No." "Maybe." "Don't know." "Last to
week."). 3 Severe poverty of speech. Answers are rarely more than a few words in length and never include additional information. Some questions left unanswered. 4 Extreme poverty of speech. Patient is essentially mute.
Ratings for severity, since they tend to be quantitative, are based standard 45-minute interview, and appropriate corrections are made if the actual evaluation interview is longer or shorter. In addition to the 18 subtypes of thought disorder, the scale also contains instructions for making a global rating. Weighted for the global rating is .89. This scale was used to evaluate the subtypes and severity of thought disorder occurring in a total of 113 patients, including 32 manies, 36 depressives, and 45 schizophrenics. To obtain a large sample and to maximize variance within the sample studied, these patients were drawn from three different hospitals that tended to have somewhat different populations. Twenty-four manies, 31 depressives, and 21 schizophrenics were evaluated at Iowa Psychiatric Hospital in Iowa City, an acute treatment facility with a two-week average length of stay, that treats both private and nonprivate patients. Four manies, four depressives, and 23 schizo¬ phrenics were drawn from Mt Pleasant State Hospital, an acute treatment facility for state patients with a mean length of stay of two months. A small number of patients at the Iowa City Veterans Administration (VA) Hospital were also evaluated: five manies, no depressives, and one schizophrenic. Patients from these facilities were included in the study if they met the Research Diagnostic Criteria for manic disorder, major depressive disorder, or schizophrenia (subacute, subchronic, and chronic types).' Although all the schizophrenic patients had had a period of illness greater than six months (including the prodromal period), none were chronic in the sense of having had long-term hospitalizations. The median number of hospitalizations was 2.5 for the schizophrenics, 3 for the manies, and 1 for the depressives. All of the schizophrenics, 84% of the manies, and 62% of the depressives had had prior hospitalizations. Patients were drawn from the Mt Pleasant and Iowa City populations on the basis of consecutive admissions that met the diagnostic inclusion criteria. We began using patients from the VA Hospital toward the end of the study, and there we also used consecutive admissions. We evaluated most patients during the first week after admis¬ sion. Nearly all were receiving medication at the time of evalua¬ tion, since medications are usually prescribed at these facilities as early as possible after admission to reduce length of stay. Although nearly all were receiving medication, all were still severely symptomatic, so that all manifested a full and typical manic, schizophrenic, or depressive syndrome at the time of the evaluation. In a few instances (limited exclusively to manic patients), patients who were admitted consecutively and met inclusion diagnostic criteria had responded to medication so quickly that they no longer manifested the full manic syndrome, and these patients were therefore not included in the study. The three diagnostic subgroups did not differ significantly from one another in age (F 0.158) or educational level (F 0.096). They did differ slightly on sex ratio using stringent criteria on a
=
=
< .135). The manies included 14 men and 18 women age of 33.0 and a mean education of 13.5 years. The depressives included 14 men and 22 women with a mean age of 39.7 and a mean education of 12.7 years. Twenty-seven schizophrenics were male and 18 female; this group had a mean age of 29.0 years and a mean education of 12.1 years. An attempt was made to give all subjects the Shipley Hartford test to assess vocabulary and
(2
=
with
4.000,
a mean
intelligence, but many of the schizophrenics and manies were too ill to cooperate fully with this task, often only completing a few items and then becoming restless. Consequently, these data were considered invalid and were not analyzed. The patients were interviewed and evaluated with the TLC Scale in two different ways. The first set of patients, a total of 44, were evaluated by means of tape-recorded interviews that had been collected in an earlier project. These interviews were approx¬ imately 20 minutes in length, not formally structured, and covered such areas as recent experiences, current events, interests, and
interpersonal relationships. Psychiatric symptomatology
was
not
discussed. These interviews were evaluated with the TLC Scale both by a blind rater (N.C.A.) and by the person who originally conducted the interview. Ratings tended to be quite close, but the data from the blind rater only are reported in this study. The second set of patients (19 manies, 20 depressives, and 30 schizo¬ phrenics) were interviewed for 45 minutes, using a standard structured interviewing procedure. This procedure began by asking the patient to talk about himself for five to ten minutes without interruption (except for prompting to keep him going as needed), covering such areas as his work, home, interests and hobbies, and personality traits. Thereafter, a standard series of follow-up questions was used, ranging from abstract to concrete,
open-ended to closed, and personal to impersonal. Psychiatric symptomatology was not discussed. Ratings were done at the conclusion of the interview by two raters, one of whom had been the interviewer and the other acting as observer. Although the interview was tape-recorded, the tape was not reviewed prior to making ratings. The patients were evaluated initially by a third person, who determined diagnosis, described the study to them, and obtained informed consent. Consequently, both raters were kept as blind as possible concerning diagnosis in the second portion of the study. Data from the two sets of subjects did not differ statistically significantly, and consequently the data from the two sets are pooled in this report. Data were analyzed with use of packaged programs from the Statistical Package for the Social Sciences or Statistical Analysis System. RESULTS Tables 1 and 2 contain simple descriptive data concern¬ ing the frequency and mean ratings of the various subtypes of thought disorder. The frequency counts in Table 1 are based on a rating of 1 or higher on the TLC Scale. A rating of 1 indicates a mild disorder, often as little as a single instance of the item being rated. In spite of this low threshold for positive ratings of these items, many of the subtypes occurred quite infrequently, including some that are often considered very important indicators of formal thought disorder, such as blocking, clanging, neologisms, and incoherence. Many of the subtypes occurred with nearly equal frequency in manies and schizo¬ phrenics, including tangentiality, derailment, incoherence, illogicality, and loss of goal. Nearly all manies and schizo¬ phrenics also had a global rating of 1 or greater. Some subtypes were commoner in mania, such as pressure of speech, clanging, distractible speech, or circumstantiality. Other subtypes occurred more frequently in schizophrenia, such as poverty of speech and poverty of content of speech. Although 53% of the depressed patients had a global rating of 1 or greater, this rating reflected subtypes that were either mild or consistent with other depressive symptoms, such as poverty of speech, poverty of content of speech, tangentiality, circumstantiality, or loss of goal. Table 2 indicates more clearly that most subtypes of thought disorder occurred in a mild form or did not occur at all in depressed patients. The transformed mean global
Downloaded From: http://archpsyc.jamanetwork.com/ by a Yale University User on 05/15/2015
Table
1.—Frequency of Types of Thought in Psychiatric Patients* No.
Derailment Incoherence
Illogicality Clanging Neologisms Word
Subsets!
Schizophrenics ( = 45) 13(29)
6(19) 23(72) 10(31) 11(34)
6(17) 2(6) 0(0) 9(25)
18(56) 5(16) 8(25) 3(9) 1(3)
5(14) 0(0) 0(0) 0(0) 0(0)
18(40) 12(27) 1(2) 16(36) 25(56) 7(16) 12(27) 0(0) 1(2)
1(3) 8(25) 14(44) 11(34) 1(3) 1(3) 2(6) 7(22) 28(88)
0(0) 11(31) 6(17) 2(6) 0(0) 2(6) 1(3) 4(11) 19(53)
32) 2(6) =
Circumstantiality Loss of goal Perseveration Echolalia
Blocking Stilted speech Self-reference Global rating on
ratings
of >1
on
_
Perseveration Global rating
Mean
Poverty of speech Poverty of content of speech Pressure of speech Distractible speech
Tangentiality Derailment Incoherence
Illogicality Clanging Neologisms Word
approxima¬
tions
Circumstantiality Loss of goal Perseveration Echolalia
Blocking Stilted Speech Self-reference Global rating
32)
phrenics ( = 45)
Mean
SD
3.9
8.0
4.7
8.4
2.2 14.1 3.8 4.4 14.4
4.9 1.6 6.1 6.7 15.2 7.8 7.6 4.2 1.8
2.2 0.5 0.0 2.8 1.7 0.0 0.0 0.0 0.0
5.4
7.6 3.6 0.2 5.5 12.2 3.3 4.2 0.0 0.2
10.9 6.5 1.5 8.4 13.3 9.0 8.7 0.0 1.5
1.8 9.2 10.8 8.0 5.3
0.0
0.0
4.2 3.1 0.1 0.0
0.0 0.4 7.1 1.3 0.7 0.6 0.2
0.0 2.1
1.6 18.0
4.1 1.3 0.3
0.3 4.6 8.4 5.0 0.9 0.3 1.3 3.1
20.0
2.3 0.0 5.1 4.5 0.0 0.0 0.0 0.0
4.9
0.2
6.9 7.5 2.3 0.0 2.3 1.6
6.4 1.1
1.1 7.0
3.2 7.4
0.5
*Scores were transformed to range from 0 to 40 by using
constant of 10.
a
SD
9.2 6.4
3.3 3.3 1.5 4.2 10.6
multiplication
7 in the depressed patients, as compared with 20 for the manies and 18 for the schizophrenics. Depressed patients tended to have poverty of speech and to be more circumstantial. Manic patients, on the other hand, tended to have their highest mean ratings on pressure of speech, derailment, tangentiality, illogicality, and loss of goal; they
rating is
D/M.S
.006 .042
S/D.M
4.790
.010
18.163
0.000
D,S/S,M D/M.S
6.435; for setwise
=
=
=
=
highest,
last.
relatively high mean ratings on distractible speech, circumstantiality, and perseveration. Further, the schizophrenic patients had considerably higher mean ratings on poverty of speech and poverty of content of speech. Table 3 indicates the results of an analysis of variance of some of the subtypes of thought disorder that occurred more frequently. Pressure of speech, derailment, distracti¬ ble speech, and the global rating have extremely high F ratios. The Scheffé procedure indicates that part of the variance for pressure of speech and distractible speech derives from differences between manies and schizophren¬ ics, while they tend to resemble one another and to differ from the depressives in terms of derailment and global rating. Poverty of content of speech and circumstantiality are other variables on which manies and schizophrenics tend to differ significantly, and there is a suggestion that poverty of speech might also be useful in differentiating these two groups. The three groups do not differ signifi¬ cantly on any other subtypes if a strict Bonferroni approach is used. These trends suggest that manies and schizophrenics tend to resemble one another in terms of illogicality, perseveration, and loss of goal. These data were also subjected to multivariate analysis of variance (MANOVA) using ten variables: poverty of speech, poverty of content of speech, pressure of speech, distractible speech, tangentiality, derailment, incoherence, illogicality, loss of goal, and perseveration. The generalized F value for this MANOVA was 8.495, with < .01. Two follow-up tests, which had been planned a priori, were done also had
6(13) 41(91)
Mean
=
disorder is listed first and the
2(4) 1(2)
Depressives ( = 36)
D.M/S D.S/M D.S/M
.012_D,M/M,S
=
2.5
3.1
.003 .000 .000 .213 .000 .079
=
0.6
SD
6.089 30.163 13.341 1.566 11.596 2.600
separately: for setwise a .025, critical .05, critical F 5.667. f M indicates manies ( 32); D, depressives ( 36); S, schizophren¬ ics ( 45). The virgule separates groups significantly different from one another; the group with the lowest score for the subtype of thought
Schizo¬
(
.044
*P is calculated for each F ratio
F
0(0) 2(4) 20(44) 11(24) 2(4)
Table 2.—Transformed Mean Ratings for Types of Thought Disorder*
=
3.217
Illogicality _4.602 5.309 Circumstantiality Loss of goal 3.270
TLC Scale.
Manies
Poverty of speech Poverty of content of speech Pressure of speech Distractible speech Tangentiality Derailment. Incoherence
approxima¬
tions
*Based
of Patients
Depressives ( = 36) 8(22)
Manies
( Poverty of speech Poverty of content of speech Pressure of speech Distractible speech Tangentiality
(%)
Table 3—Analysis of Variance and Scheffé Procedure for Differences Between Subsets on Types of Thought Disorder
Disorder
the MANOVA. These are summarized in Table 5. These follow-up tests can be used to compare two different ways of conceptualizing "thought disorder." The first test evalu¬ ates the concept of loose associations by giving an equal positive weighting to each of the five subtypes considered to represent various forms of associative loosening (tan¬ gentiality, derailment, incoherence, illogicality, and clang¬ ing). This follow-up test yields an extremely low F value and does not disclose any significant differences between the three groups in terms of associative loosening. An alternative approach was to conceptualize "thought disoron
Downloaded From: http://archpsyc.jamanetwork.com/ by a Yale University User on 05/15/2015
Table 4.—Variance in Manic and Schizophrenic Patients Accounted for by Discriminant Variables Entered Stepwise
(N
-
77)*
Unstandardized
Coefficients! Pressure of speech Distractible speech Derailment
.61578 .95629 .28826
% of Variance 40.251 17.815 16.457
Cumulative % 58.066 74.523
Poverty of content 12.300 86.823 of speech _.37874 6.808 Poverty of speech_.37843 _93.631 .22657 4.984 Loss of goal 98.615t *These other variables were insufficiently discriminatory and were therefore not entered into the analysis: perseveration, tangentiality, inco¬ herence, and illogicality. tConstant .88329. JDoes not add to 100% due to rounding error. =
Table
5.—Follow-up Tests of Multivariate Analysis of Variance (N 113) =
Loose associations* +1
Tangentiality Derailment Incoherence
+1
Illogicality Clanging Positive and negative formal thought disordert Poverty of speech Poverty of content of speech Pressure of speech
+ 1
Tangentiality
+1
Derailment Incoherence
+1
Illogicality •F
=
fF
=
0.4248, 9.918,
=
