Defines “mental illness.” The aggregation of the disorders of the DSMs has been the operating definition of “mental illness.” [See discussion of DSM’s efforts to define “mental illness” per se, see near the end of the document, “Some Issues Already Before The DSM-V Task Force”].
Defines for what clinicians can be held accountable in terms of knowledge and skills.
Defines responsibility for the public psychiatric sector.
Defines reimbursibility for the private psychiatric sector.
Values.
To communicate [e.g., “bipolar disorder, mixed type“]
To determine treatment [e.g., “FDA approved for schizophrenia”]
To explicate the “truth,” especially the “cause” [e.g., “dementia due to Huntington’s disease”]
To avoid stigmatizing the person with the illness, the environment or the family.
To communicate or to determine treatment?
Does the diagnostic system achieve what it is intended to achieve?
1] Communicative validity = descriptive validity
A] To the pt
B] To others working with the pt
C] To the profession in order to increase the knowledge about psychiatric illnesses, their treatment, and their prevention
2] Treatment validity = predictive validity
Table comparing features which increase ability to communicate and features which increase ability to predict:
TO COMMUNICATE
TO TREAT
Simplicity
Complicated
Constrictive
Flexible
Evaluate consensually
Evaluate empirically
The DSMs goal has been “to communicate.” Of the five DSMs, only DSM-IIIR had a treatment response as part of one of its criteria. [Criteria for Melancholic Type, included: “previous good response to specific and adequate somatic antidepressant therapy, e.g., tricyclics, ECT, MAOI, lithium.”]
DSMs have had difficulty being relevant to treatment:
Kupfer, First and Regier:
“With regard to treatment, lack of treatment specificity is the rule rather than the exception.” “The efficacy of many psychotropic meds cut across the DSM-defined categories. For example, the SSRIs have been demonstrated to be efficacious in a wide variety of disorders, described in many sections of DSM.” [They list eleven different DSM disorders to which SSRIs might be useful.]
A diagnostic system for patients whose diagnosis has no known validity and addressed only medication selection could probably get by with 13 “diagnostic” categories: 1. Agitation. 2. Attention deficit/hyperactive. 3. Binge eating. 4. Cognitive decline. 5. Dysphoria/anxiety. 5a. Anxious, not dysphoric. 5b. Dysphoric, not anxious. 6. Enuresis. 7. Excessive sleepiness. 8. Impulse dyscontrol. 9. Insomnia. 10. Mania/psychosis. 10a. Psychotic and suicidal. 11. Mood swings. 12. Suicidal. 13. Tourette’s.
Validities in DSM-IV-TR.
Validity comes in a number of versions, including:
1 – Etiology – see infra*
2 - Event/environmental – stress/trauma**
3 - Genetic/family – examples are illnesses shared with the rest of medicine, e.g., Huntington’s.
4 – Biological marker – polysomnography
5 – Psychological test finding -- IQ
Not in DSM-IV-TR:
1 – Prognosis – none, but effort to reach prognosis in setting time requirements, e.g., schizophrenia, adjustment disorder
2 – Treatment effects – none
** Criteria sets, unlike the rest of medicine, may reflect on the environment in which the individual exists. Two examples: ADHD’s signs and symptoms have to be seen in two different settings. Phobias are not a Disorder unless they are in a setting in which they are relevant to the person.
*DSM-IV dxs having some validity, some etiological suggestion:
Dx related to substances = 124
Dx related to illnesses shared with the rest of medicine = 36 [obviously many more not mentioned in DSM-IV-TR]
Dx related to stress/trauma = 9
Dx related to season = 1
Dx related to post-partum time = 1
Total having some etiological elements: 171 [about half of the DSM-IV-TR]
Review of DSM-IV-TR valid diagnoses suggests that etiological-based dx will have anxiety, mood and psychotic features that they will not fit into the organization of psychopathology of DSM-IV-TR’s not-yet-validated section.
Important reflections on validity.
Steve Hyman:
“Despite these successes [of the DSMs], there are clear problems and unresolved controversies related to DSM-IV-TR, the most recent version of DSM. If a relative strength of DSM is its focus on reliability, a fundamental weakness lies in the problems related to validity. Not only persisting but looming larger is the question of whether DSM-IV-TR truly carves nature at the joints – that is, whether the entities described in the manual are truly ‘natural kinds’ and not arbitrary chimeras.”
“In reifying DSM-IV-TR diagnoses, one increases the risk that science will get stuck, and the very studies that are needed to better define phenotypes are held back.”
“Except for IQ tests to diagnose mental retardation and polysomnography to diagnose sleep disorders [polysomnography was inexplicitly excluded from DSM-IV-TR criteria sets], diagnostic tests for mental disorders do not yet exists.”
“Overall, there is evidence that the current diagnoses and their corresponding sets of criteria fall short of mapping nature.”
“It will be important to avoid premature inclusion of genetic or neurobiological findings in the DSM, no matter how interesting they are, if they are not adequately replicated or if their phenotypes cannot be established with clarity. At the same time, a slavish adherence to the current classification system would impede progress in research that is investigating the etiology of mental disorders and identifying new treatment for them”
Reliability = consistent results across clinicians and across time.
Reliability is a worthy goal of the DSMs to facilitate communications:
-- With patients -- with others working with or living with the patient
-- Within the profession to increase knowledge
-- With health care policy-makers.
Coverage = the applicability of the classification system.
DSM’s goal of coverage is to provide a diagnostic term for all the people who policy-makers and practitioners want to include as having a mental/psychiatric disorder/illness.
Table comparing achieving reliability and coverage is on the next page. The APA wants coverage of 100% applicability of the DSMs. It wants reliability of consistent results across clinicians and across time.
ASPECT
COVERAGE
RELIABILITY
Rules
Flexible
Precise
Terms are
Vague
Explicit
Information needed to make a dx
Broad allowances acceptable
Very consistent quantity of information
Diagnostician’s training
Broad allowances as to training
Very similar training expected
Interview setting
Broad allowances as to setting
Very similar interview setting
Resolution of coverage – reliability conflict.
DSMs achieve both coverage and reliability through:
1. Specific criteria sets to increase reliability.
2. Allowing dual or more diagnoses.
3. Having NOS options throughout.
A reliability issue: Combinations with criteria sets.
The potential combinations within a given psychiatric dx can be enormous. While many of DSM's criteria set have the characteristics of a medical syndrome where “and” is used, some of the DSM criteria sets with “or” can have combinations that can be huge. The number of possible bipolar disorders is in the billions. Borderline Disorder has 256 combinations. Reliability would be markedly enhanced by defining Borderline Disorder with just 3 requirements:
– 1. A pattern of disruptively unstable interpersonal relationships,
– 2. A pattern of self-damaging impulsive behavior, AND
– 3. A pattern of troublesome affective instability.
[Other frequent signs would go into “Associate Features” part of the DSM text]
Number of Disorders – 374. Additionally there are 23 V-Codes
Number of Codes – 256. 118 of these repeat to be compatible with ICD coding.
K. Avoiding stigmatizing involves conceptualizing patients as separate from their mental illness.
-- DSM III’s criteria set created a climate of the person is not the illness.
-- DSM IV specifically calls for use of “person with schizophrenia” rather than “schizophrenic.”
-- Personality disorders, however, still imply a reflection on the person in the eyes of many, so Washington Psychiatric Society and the Maryland Psychiatric Society have passed motions, 2002, asking the APA to abolish “personality” in titles, e.g. “narcissistic personality disorder” should be replaced by “narcissistic disorder.”
L. Avoiding stigmatizing/blaming the patient or family.
-- Prediction: In the near future, we will not see any DSM suggestion that parenting can lead to a specific psychiatric disorder.
-- Proposed "Self-defeating Disorders" was removed even from Disorders to be Studied because of fear of blaming the victim of abuse.
-- Relegates “Premenstrual Dysphoric Disorder” to Disorders to be Studied to avoid stigmatizing women, even though in ICD-9-CM.
M. Avoid a system in which everyone has or has had a mental illness.
-- Categorical conceptualization = syndrome with an identifiable separation from normalcy.
-- In many criteria sets include: “ . . . cause clinically significant distress or impairment in social, occupational, or other important areas of functioning.” [Majority of NOSs, however, don’t carry this stipulation.]
N. Organizing Options:
How should psychopathology be organized?
Possibilities:
-- Etiology, e.g., chronic over-use of alcohol is associated with 16 DSM-IV disorders.
-- Shared psychopathology, shared symptoms/signs, e.g., mood disorders
-- Shared non-psychopathological characteristics, age, gender, race, and culture, e.g., Disorders of infancy, children, adolescents, adults, and aged.
-- Shared epidemiology togetherness [see Spectra proposals in DSM-V-issues handout], not a DSM-IV organizing principle
-- Shared familiar findings, not a DSM-IV organizing principle, e.g., combine anxiety and depressive disorders.
-- Shared neuroscience findings, not yet a DSM-IV organizing principal
--- Genetics
--- Neural pathways
--- Phenotypes
-- Shared prognosis, shared course of illness, not a major DSM-IV organizing principle
-- Shared treatment response, not a DSM-IV organizing principle.
-- Others??
O. Categorical v. Dimensional.
Categorical = state that can be qualitative separate from normal.
Dimension = quantitatively traits, traits that are continuous with normal.
The dimensional approach would enhance:
exactness
individualization
simplicity
Communication, no NOS dx needed.
reliability
coverage [depends on threshold rules]
The dimensional approach would appear to be devastatingly disruptive break with our past.
