THE PROCESS OF DIAGNOSIS:ADVANTAGES OF DIAGNOSIS, DESCRIPTION

<< CAUSES OF MENTAL ILLNESOVERVIEW OF ETIOLOGY:PANDAS

THE CONCEPT OF PSYCHOLOGICAL ASSESSMENT IN CLINICAL PSYCHOLOGY >>

Clinical Psychology (PSY401)

LESSON 14

THE PROCESS OF DIAGNOSIS

THEIMPORTANCE OF DIAGNOSIS

Whyshould we use mentaldisorder diagnoses? Diagnosis is a type of expert-levelcategorization.

Categorization is essential to our survivalbecause it allows us to makeimportant distinctions(for

example, a mild cold versus viralpneumonia, a malignant versus a benign tumor). The diagnosis of

mental disorders is an expert level of categorization used by mentalhealth professionals that enables us

to make important distinctions(for example, schizophrenia versusbipolar disorder withpsychotic

features).

ADVANTAGES OF DIAGNOSIS

Thereare at least four majoradvantages of diagnosis. First, and perhaps most important, a primary

function of diagnosis is communication. A wealth of information can be conveyed in a single diagnostic

term. For example, if a patientwith a diagnosis of paranoid schizophrenia is referred to a psychologist,

immediately,without knowing anythingelse about the patient, a symptom pattern will come to mind

(delusions, auditory hallucinations, severesocial/occupational dysfunction, andcontinuous signs of the

illnessfor at least 6 months). Diagnosiscan be thought of as"verbalshorthand" for representing features

of a particular mental disorder.Using standardized diagnostic criteria(such as those that appear in the

DSM-IV)ensures some degree of comparability with regard to mentaldisorder features among patients

diagnosed in the same area or region.

Diagnosticsystems for mental disorders are especially useful forcommunication becausethese

classificatorysystems are largelydescriptive. That is, behaviorsand symptoms that are characteristic of

the various disorders are presentedwithout any reference to theories regarding their causes. As a result,

a diagnostician of nearly anytheoretical persuasion canuse them. If every psychologistused a different,

theoreticallybased system of classification, a great number of communication problems wouldlikely

result.

Second, the use of diagnoses enables and promotes empirical research in psychopathology. Clinical

psychologists define experimental groups in terms of individuals' diagnostic features, thusallowing

comparisons between groups with regard to personality features, psychological test performance, or

performance on an experimental task. Further, the waydiagnostic constructs aredefined and described

willstimulate research on the disorders' individual criteria, on alternativecriteria sets, and on the

comorbidity (co-occurrence) between disorders.

Third, and in a related vein, researchinto the etiology, or causes, of abnormal behavior would be almost

impossible to conduct without a standardized diagnosticsystem. In order to investigate the importance

of potential etiological factors for a given psychopathological syndrome, we must first assign subjects to

groups whose members sharediagnostic features. For example, several years ago it was hypothesized

that the experience of childhood sexual abusemay predispose individuals to develop features of

border-linepersonality disorder (BPD).The first empirical attempts to evaluate the veracity of this

hypothesis involved assessing the prevalence of childhood sexual abuse in well-defined groups of

subjectswith borderline personalitydisorder as well as in non-borderlinepsychiatric controls.These

initialstudies indicated thatchildhood sexual abuse does occur quite frequently in BPDindividuals and

thatthese rates aresignificantly higher thanthose found in patients withother (non-BPD)mental

disorderdiagnoses. Before we couldreach these types of conclusions, there had to be a reliable and

systematic method of assigning subjects to the BPDcategory.

Finally,diagnoses are importantbecause, at least in theory,they may suggest whichmode of treatment

is most likely to be effective.Indeed, this is a general goal of a classification system formental

107

Clinical Psychology (PSY401)

disorders (Blashfield & Draguns, 1976). As Blashfield and Draguns (1976, p.148)stated, "The final

decision on the value of a psychiatricclassification for predictionrests on an empirical evaluation of the

utility of classification for treatment decisions." For example, a diagnosis of schizophrenia suggests to

us that the administration of an antipsychotic medication is more likely to be effective than is a course

of psychoanalytic psychotherapy. However, it is important to note one thing in passing. Although, in

theory, the linkage between diagnosis and treatment wouldseem to justify the timeinvolved in

diagnosticassessment, often several treatmentsappear to be equally effectivefor an individualdisorder.

In summary, diagnosis and classification of psychopathology serves manyuseful functions.Whether

theyare researchers or practitioners,contemporary clinical psychologists usesome form of diagnostic

scheme in their work.

At this point, we turn to a brief description of classificationsystems that have been used to diagnose

mental disorders over the years, and then we examine in more detail the features of the diagnostic

classificationsystem that is used mostfrequently in the United States, the DSM-IV.

EARLY CLASSIFICATION SYSTEMS

Classificationsystems for mental disorders have proliferated for many years.For example, the earliest

reference to a depressive syndrome appeared as far back as 2600 B.C. Since thattime, both the number

of and breadth of classification systems have increased.

To bring some measure of orderout of this chaos, the Congress of Mental Science adopted a single

classificationsystem in 1889 in Paris.More recent attempts can be traced to the WorldHealth

Organization and its 1948 InternationalStatistical Classification of Diseases,Injuries, and Causes of

Death,which included a classification of abnormal behavior.

In 1952, the American PsychiatricAssociation published itsown classification system in the Diagnostic

and Statistical Manual, and this manual contained a glossary describing each of the diagnostic

categoriesthat were included. Thisfirst edition, known as DSM-I, was followed by revisions in 1968

(DSM-II),1980 (DSM-III), and 1987(DSM-III-R).

Presently, the most widely usedclassification system is the previouslymentioned AmericanPsychiatric

Association'sDiagnostic and Statistical Manual of Mental Disorders, 4th edition(DSM-IV), which

appeared in 1994. All of these manualsare embodiments of Emil Kraepelin'sefforts in the late

nineteenthcentury.

Themost revolutionary changes in the diagnostic system were introduced in DSM-III, published in

1980.These changes included the use of explicit diagnosticcriteria for mental disorders, a multiaxial

system of diagnosis, a descriptive approach to diagnosis that attempted to be neutral with regard to

theories of etiology, and a greater emphasis on the clinical utility of the diagnosticsystem. Because

theseinnovations have been retained in subsequent editions of the DSM(DSM-III-R and DSM-IV),

theseare described in the following section.

DSM-IV

Thefourth edition of the Diagnostic and Statistical Manual of Mental Disorders (DSM-IV) was

published in 1994. Revisions to the previousdiagnostic manual (DSM-III-R) were guided by a three-

stageempirical process.

First,150 comprehensive reviews of the literature on important diagnosticissues were conducted. These

literaturereviews were both systematic and thorough. Results from thesereviews led to

108

Clinical Psychology (PSY401)

recommendations for revisions and served to document the rationale and empirical supportfor the

changesmade in DSM-IV.

Second, 40 major re-analyses of existingdata sets were completed in cases where the literature reviews

couldnot adequately resolve the targeted diagnostic issue.

Third, 12 DSM-IV field trials were conducted in order to assess the clinicalutility and predictivepower

of alternative criteria setsfor selected disorders (forexample, antisocial personalitydisorder). In

summary, the changes made in DSM-IV were based on empirical data to a much greater extent thanwas

true in previous editions of the DSM.

A complete DSM-IV diagnostic evaluation is a multi-axial assessment. Clients or patients are evaluated

alongfive axes, or domains of information.Each of these axes/domainsshould aid in treatment planning

and prediction of outcome.

Axis I is used to indicate the presence of any of the clinical disorders or otherrelevant conditions,with

the exception of the personality disorders and mental retardation. Thesetwo classes of diagnosesare

coded on Axis Il.

Axis III is used to highlight anycurrent medical conditionthat may be relevant to the conceptualization

or treatment of an individual's Axis I or Axis II clinical disorder. Psychosocial and environmental

problems relevant to diagnosis, treatment, and prognosis areindicated on Axis IV.

Finally, a quantitative estimate (1 to 100) of an individual's overall level of functioning is provided on

Axis V. Each of the five axescontributes important informationabout the patient, and togetherthey

provide a fairly comprehensive description of the patient's major problems, stressors, and level of

functioning.

THECASE OF MICHELLE M

Michelle M. was a 23-year-old woman whowas admitted to an inpatientunit at a hospital following her

sixthsuicide attempt in two years.She told her ex-boyfriend(who had broken up with her a week

earlier)that she had swallowed a bottle of aspirin, and he rushed her to the local emergency room.

Michelle had a five-year history of multipledepressive symptoms thatnever abated; however, these had

notbeen severe enough to necessitate hospitalization or treatment. They included dysphoricmood, poor

appetite,low self-esteem, poorconcentration, and feelings of hopelessness.

In addition, Michelle had a history of a number of rather severe problems that hadbeen present since

her teenage years. First, shehad great difficulty controlling her emotions. She was prone to become

intenselydysphoric, irritable, or anxious almost at a moment's notice. These intense negative affect

states were often unpredictable and, althoughfrequent, rarely lasted more than four or five hours.

Michellealso reported a long history of impulsive behaviors, including polysubstance abuse and binge

eating.Her anger was unpredictable and quite intense. For example,she once used a hammer to literally

smash a wall to pieces following a bad grade on a test.

Michelle'srelationships with her friends,boyfriends, and parents were intense and unstable. People who

spenttime with her frequentlycomplained that she wouldoften be angry with them and devalue them

for no apparent reason. She alsoconstantly reported an intense fearthat others (including her parents)

might abandon her. For example,she once clutched a friend'sleg and was dragged out the door to her

friend'scar while Michelle tried to convince the friend to stayfor dinner. In addition, she had attempted

to leave home and attend college in nearby cities on fouroccasions. Each time, shereturned home

within a few weeks. Prior to her hospital admission, her words to her ex-boyfriendover the telephone

were, "I want to end it all. No one lovesme."

109

Clinical Psychology (PSY401)

TheDSM-IV diagnostic evaluationfor Michelle M. is shown here:

Diagnostic Evaluation: Michelle M.

Axis I: 300.4

DysthymicDisorder early onset

305.00Alcohol Abuse

305.20 Cannabis Abuse

305.60 Cocaine Abuse

305.30Hallucinogen Abuse

Axis II: 301.83 Borderline PersonalityDisorder

(PRINCIPALDIAGNOSIS)

Axis III:

none

Axis IV:

Problems with primary support groupEducational problems

Axis V: GAF = 20 (Current)

Several features of this diagnostic formulationare noteworthy First,Michelle has receivedmultiple

diagnoses on Axis I. This is allowed, and even encouraged, in the DSM-IV system because the goal is to

describe the client's problems comprehensively.

Second,note that her border-linepersonality disorder (BPD) diagnosis on Axis II is considered to be the

principal diagnosis. This means thatthis condition is chiefly responsible for her admission to the

hospital and may be the focus of treatment.

Finally, her Global Assessment of Functioning(GAF) score on Axis V indicates serious impairment - in

thiscase, a danger of hurtingherself.

GENERALISSUES IN CLASSIFICATION

We have briefly described the DSM-lV to give the reader a general idea of what psychiatric

classificationentails. However, it is important to examine a number of broad issuesrelated to

classification in general, and to the DSM-IVspecifically. The eightmajor issues in classificationare

discussedbelow.

CATEGORIESVERSUS DIMENSIONS

Essentially, the mental disorder categoriesrepresent a typology. Basedupon certain presenting

symptoms or upon a particular history of symptoms, the patient is placed in a category. This approach

has several potential limitations. First, in too many instances, it is easy to confuse suchcategorization

withexplanation. If one is not careful, there is a tendency to think "This patient is experiencing

obsessionsbecause she has obsessive-compulsive disorder" or "This person is acting psychoticbecause

he has schizophrenia." Whenthis kind of thinkingoccurs, explanation has been supplanted by a circular

form of description.

In addition, abnormal behavior is not qualitatively differentfrom so-called normal behavior.Rather,

theseare endpoints of a continuousdimension. The difference between so-called normal behaviorand

psychoticbehavior; for example, is one of degree rather than kind. Yetmental disorder diagnoses in

terms of categories imply thatindividuals either have the disorder in question or they do not.This all-or-

110

Clinical Psychology (PSY401)

nothingtype of thinking may be at odds with what we knowabout how symptoms of psychopathology

aredistributed in the population.

Forexample, a categorical model of borderline personality disorder(BPD), as presented in the DSM-IV

(that is, present versus absent),may not be appropriatebecause individuals differonly with respect to

howmany BPD symptoms theyexhibit (a quantitative difference). In other words, the categoricalmodel

maymisrepresent the true nature of the borderline construct. In fact, there may be relatively few

diagnosticconstructs that are trulycategorical in nature.

BASES OF CATEGORIZATION

In order to classify psychiatric patients, one must use a wideassortment of methods and principles. In

somecases, patients are classified almost solely on the basis of theircurrent behavior or presenting

symptoms. In other cases, the judgment is made almost entirely on the basis of history. In the case of

majordepression, for example, one individual may be diagnosed on the basis of a diagnostic interview

conducted by a clinician; another may be classifiedbecause of a laboratory result,such as a "positive"

dexamethasonesuppression test (DST);still another may be diagnosed as a result of scores on a self-

reportmeasure of depression. Laboratoryresults provide the basisfor some diagnoses of cognitive

disorders (for example, vascular dementia),whereas other cognitivedisorder diagnoses (such as

delirium)are determined solely by behavioral observation. Thus, the diagnostic enterprise may be quite

complicatedfor the clinician, requiringboth knowledge of and access to a wide variety of diagnostic

techniques. A major implication is that membership in any one diagnostic category is likely to be

heterogeneousbecause there are multiplebases for a diagnosis.

PRAGMATICS OF CLASSIFICATION

Psychiatricclassification has alwaysbeen accompanied by a certaindegree of appeal to medical

authority.But there is a concurrent democratic aspect to the system that is quitepuzzling. Forexample,

psychiatryfor many years regarded homosexuality as a disease to be cured through psychiatric

intervention. As a result of society's changingattitudes and other validpsychological reasons,

homosexualitywas dropped from the DSMsystem and is now regarded as an alternate lifestyle. Only

when homosexual individuals are disturbed by their sexual orientation or wish to change it do we

encounter homosexuality in the DSM-IV (as an example under the category"sexual disorder not

otherwisespecified"). The issue here is not whether this decisionwas valid or not. Theissue is how the

decision to drop homosexuality from the DSM system was made.The demise of homosexuality as a

diseaseentity occurred through a vote of the psychiatric membership.

Thisexample also serves as a reminder that classificationsystems such as the DSMare crafted by

committees. The members of such committees represent varying scientific,theoretical, professional, and

even economic constituencies. Consequently, the finalclassification product adopted mayrepresent a

political document that reflects compromises thatwill make it acceptable to a heterogeneous

professionalclientele.

DESCRIPTION

Withoutdoubt, the DSM-IV providesthorough descriptions of the diagnostic categories. The DSM also

providesadditional information foreach diagnosis, including the age of onset, course, prevalence,

complications,family patterns, cultural considerations, associated descriptive featuresand mental

disorders, and associated laboratoryfindings. All this descriptivedetail should enhance the system's

reliabilityand validity.

RELIABILITY

A scheme that cannot establish itsreliability has serious problems. In this context, reliability refers to

the consistency of diagnostic judgments across raters. One of the major changes seen in DSM-III (the

111

Clinical Psychology (PSY401)

inclusion of specific and objectivecriteria for each disorder)reflected an attempt to increase the

reliability of the diagnostic system. If Psychologist A and Psychologist B both observe the same patient

but cannot agree on the diagnosis, then boththeir diagnoses are uselessbecause we do not knowwhich

to accept. This is the verysituation that plagued the American diagnostic systemsfor many years.

Forexample, an early studyillustrating the unreliability of previous diagnostic systemswas carried out

by Beck, Ward, Mendelson, Mock, and Erbaugh (1962). Twodifferent psychiatrists eachinterviewed

the same 153 newly admittedpsychiatric patients. Overall agreement among these psychiatrists was

only54%. Some of the disagreements in diagnosis seemed to stem from inconsistencies in the

information patients presented to the psychiatrists. Forexample, Patient A may have been relatively

open with Psychiatrist F, but less so with Psychiatrist G. But much of the unreliability problemseemed

to lie with the diagnosticiansand/or the diagnostic systemitself.

Certainpragmatic factors can alsoreduce reliability acrossdiagnosticians. Sometimes it happensthat a

giveninstitution will not admit patients who carry a certain diagnosis. Yet a mental healthprofessional

mayfeel strongly that the patient could benefit from admission (or perhaps hasnowhere else to go).

Whatshould be done? The"humanitarian" choice often seems to be to alter a diagnosis, or at least to

"fudge" a bit. The patient withalcohol dependence suddenly is diagnosed with something else.

Similarly, an insurance company may reimburse a clinicfor the treatment of patients with one diagnosis

butnot another. Or perhaps one diagnosis permits six therapy visitsbut another allows as many as 15

sessions.Therefore, a diagnosis may be intentionally or unintentionallymanipulated.

Theseexamples may lead us to believe that diagnosticunreliability is the rule and not the exception.

However,Meehl (1977), for example, feels that psychiatric diagnosis is notnearly as unreliable as it is

madeout to be. Specifically,Meehl argues that if we confine ourselves to major diagnosticcategories,

requireadequate clinical exposure to the patient, and study well-trainedclinicians who take diagnosis

seriously,then inter-clinician agreementwill reach acceptablelevels.

Thefield of psychopathology has begun to address these concerns aboutreliability by developing

structured diagnostic interviews thatessentially "force" diagnosticians to assess individuals for the

specificDSM criteria that appear in the diagnostic manual. Forexample, there are now several

structured interviews that assessfeatures of Axis I disorders and a number of structured interviews for

Axis II disorders exist as well.Interestingly, the overall level of diagnostic reliability reported in

empiricalstudies has increasedgreatly following the introduction of these structured interviews. It is

clear that adhering to the structure and format of these interviews has led to a significant increase in

diagnosticreliability.

However, even with the use of structured interviews, reliability is notequally good across allcategories.

Thepresence versus absence of some disorders (for example, generalized anxiety disorder) may be

particularlydifficult to judge.

Further, there is some question as to whether or not busy clinicians willdevote the time and effort

necessary to systematically evaluate the relevantdiagnostic criteria. Reliabilitycoefficients neverseem

to be as high in routine, everydaywork settings as they are in structured research studies.

VALIDITY

Reliabilitywill directly affect the validity of a diagnostic system. As long as diagnosticians fail to agree

upon the proper classification of patients, we cannot demonstrate that the classificationsystem has

meaningful correlates - that is, has validity.Important correlates include prognosis, treatment outcome,

wardmanagement, etiology, and so on.And without predictivevalidity, classification becomes an

intellectualexercise devoid of anyreally important utility.However, if we can demonstratethat

categorization accurately indicates etiology, course of illness, or preferred kinds of treatment, then a

validbasis for its usehas been established.

112

Clinical Psychology (PSY401)

Thepredominant method for establishing the validity of a diagnostic construct wasout-lined in a classic

article by Robins and Guze (1970). Theyproposed that establishing the diagnostic validity of a

syndrome is a five-stage process: (1)clinical description, including a description of characteristic

featuresbeyond the disorder's symptoms (such as demographic features); (2) laboratorystudies

(includingpsychological tests) to identifymeaningful correlates of the diagnosis; (3)delimitation from

other disorders to ensure some degree of homogeneity among diagnostic members;(4) follow-up studies

to assess the test-retest reliability of a diagnosis; and (5) family studies to demonstrate that the proposed

disordertends to run in families, suggesting a hereditary component to the disorder.This particular five-

stage method for establishing diagnosticvalidity remains quiteinfluential even today. In fact,most

contemporaryresearch in psychopathology representsone or more of the validation stagesoutlined by

Robins and Guze.

BIAS

Ideally, a classification system willnot be biased with respect to how diagnoses areassigned to

individualswho have different backgrounds (forexample, different gender,race, or SES). Thevalidity

andutility of a classification systemwould be called intoquestion if the same cluster of behaviors

resulted in a diagnosis for one individual butnot for another individual.The two areas of potentialbias

that have received the most attentionare sex bias andracial bias.

Somecritics have attacked the DSM system as a male-centered device that overestimates pathology in

women,others deny this charge.Widiger and Spitzer (1991) have presented a useful conceptual analysis

of what constitutes sex bias in a diagnostic system. They argue that previous attempts to demonstrate

diagnosticsex bias have been bothconceptually and methodologicallyflawed. Further, some of the

findings of earlier studies have beengrossly misinterpreted and misunderstood.

Widiger and Spitzer note thatdifferential sex prevalence for a disorder does not in and of itself

demonstratediagnostic sex bias because,for example, it is conceivablethat biological factors or cultural

factors may make it more likely thatmen (or women) willexhibit the criteria for a certain diagnosis. For

example,antisocial personality disorder is diagnosed much more frequently in menthan in women, but

thismay be the result of biologicaldifferences (such as testosterone) or other factors that influence the

twogenders differentially (such as societal expectations for aggressiveness in men).

However,Widiger and Spitzer didpresent evidence suggesting thatclinicians may be biased in the way

theyapply diagnoses to menversus women, even in cases where the symptoms presented by menand

women were exactly the same! Althoughthis suggests that there may be some bias in the wayclinicians

interpret the diagnostic criteria (that is, clinicians may exhibit sex bias), it does not indicatesex bias

within the diagnostic criteria. Theseresults suggest the need forbetter training of diagnosticiansrather

than an over-haul of the diagnosticcriteria.

COVERAGE

Withclose to 400 possible diagnoses,DSM-IV cannot be faulted forbeing too limited in its coverage of

possible diagnostic conditions. It is likelythat most conditions thatbring individuals in forpsychiatric

or psychological treatment could be classified within the DSM-IVsystem. However, some mayfeel that

DSM-IVerrs in the opposite direction - that its scope is toobroad. For example, a host of childhood

developmental disorders are included as mental disorders. The child who is dyslexic, has speech

problems such as stuttering, or has great difficulties with arithmetic is given a DSM-IV diagnosis. Many

question the appropriateness or benefit of labeling these conditions as mental disorders.

ADDITIONALCONCERNS

Although the previously described difficultiesare real and fairly obvious, a number of indirect or subtle

problems arise through the acceptanceand use of diagnosticclassification systems. Forexample,

113

Clinical Psychology (PSY401)

classificationstend to create the impression thatmental disorders exist per se.Such terms as disorder,

symptom, condition, and sufferingfrom suggest that the patient is the victim of a diseaseprocess. The

language of the system can eventuallylead even astute observerstoward a view thatinterprets learned

reactions or person-environment encounters as disease processes.

In addition, if we are notcareful, we may come to feelthat classifying people is more satisfying than

trying to relieve their problems. As we shallsee later, therapy can be an uncertain, time-consuming

processthat is often fraught withfailure. But pigeonholingcan be immediately rewarding: it provides a

sense of closure to the classifier. Likesolving crossword puzzles, it may relieve tension withouthaving

anylong-term positive socialsignificance.

Thesystem likewise caters to the public's desire to regard problems in living as medical problems that

can be dealt with simply andeasily by a pill, an injection, or a scalpel. Unfortunately, however,learning

to solve psychological problems is hardwork. The easier approach is to adopt a passive, dependent

posture in which the patient is relieved of psychological pain by an omniscientdoctor. Although such a

viewmay be serviceable in dealing withstrictly medical problems, it hasdubious value at best in

confronting the psychosocial problems of living.

A final indirect problem is that diagnosis can be harmful or even stigmatizing to the person who is

labeled. In our society, diagnosis mayclose doors rather than open them for patients and ex-patients.

Toooften, diagnosis seems to obscure the real person; observers seelabels, not the real peoplebehind

them. Thus, labels can damage relationships,prevent people from beinghired or promoted, and, in

extreme cases, even result in a loss of civil rights. Labels can even encourage some people to capitulate

and assume the role of a "sick"person.

114

Table of Contents: