Friday, February 12, 2010

Critque of proposed DSM5 intellectual disability criteria: Guest post by Dr. Dale Watson


Without question, the DSM-5 Proposed Draft Revision document has been generating considerable chatter among psychologists.  With regard to Atkins cases, the proposed definition of intellectual disability (ID) is no exception....emails and listservs have been busy debating and critiquing the ID proposed criteria. 

Dr. Dale Watson has set out a well-written set of concerns and issues in the guest blog post below---which is reproduced "as is" from Dr. Watson.  Kudos to Dale for providing ICDP with his perspective.

Dr. Dale Watson's critique of the proposed DSM-V ID criteria follows:

The DSM-5 Proposed Draft Revisions to the Criterion sets for Mental Disorders have recently become available.  The proposed criteria for the diagnosis of Intellectual Disability retain the three-pronged model of diagnosis used by both the DSM-IV and the AAIDD.  However, the revised language, though more precise in some ways, is also potentially problematic for a number of reasons and requires further clarification.  The following critique outlines concerns regarding the revised language and is a request for further clarification and/or specificity in the diagnostic language.

The first prong of the revision appears initially to improve the specificity of the IQ requirements stating, “Current intellectual deficits of two or more standard deviations below the population mean, which generally translates into performance in the lowest 3% of a person’s age and cultural group, or an IQ of 70 or below.”  Certainly using “standard deviations below the population mean” adds a degree of precision and perhaps allows for consideration of “Flynn Effect” changes in the population mean.  However, in an Atkins context, this language also appears to foreclose sole reliance on historical test scores in establishing the diagnosis in that it requires “current intellectual deficits.”  In addition, by eliminating the DSM-IV “IQ of approximately 70 or below” there must be some concern that this proposal establishes a “bright-line” cutoff of “70 or below” for the diagnosis of an Intellectual Disability.  Despite the rationale provided by the Work Group that the proposed criteria add “rigor to wording regarding psychometrics [with] (no change in cut –off)” one must be concerned that this is, in effect, a change in the cutoff.  In justifying the changes in the coding procedure, the rationale indicates, “Inaccuracy of testing no longer a factor.”  Does this mean that eliminating the word “approximately,” as used in DSM-IV and by the AAIDD, also eliminates consideration of the Standard Error of Measurement?  Does this mean, as is the practice in a number of death-penalty states, that it is no longer “possible to diagnose Mental Retardation with IQ scores between 71 and 75…”  (DSM-IV, p. 48).  If that were the case, the proposed criteria would not be consistent with clinical practice nor would it, by ignoring the standard error or measurement, be “adding [psychometric] rigor.”

The revision language also acknowledges the importance of cultural sensitivity, which, on the face of it, should not be objectionable.  However, by stating that the IQ requirement “translates into performance in the lowest 3% of a person’s age and cultural group…” there is a risk that this will be used as a rationale to establish subgroup norms rather than relying, as indicated, on the “population mean.”  Certainly we have seen attempts by some psychologists to use membership in a presumed cultural group to inflate IQ scores thus making legitimately intellectually disabled individuals eligible for the death penalty.  These attempts have included the use of Heaton’s WAIS-III demographically adjusted norms and what have been termed “IQ-Quality” scores, both of which inflate the obtained IQ scores based upon subgroup membership to establish that individuals do not have an Intellectual Disability.  IQs, by definition, must reference population rather than subgroup standards.

There is also either ambiguity or a substantially increased demand for deficits in adaptive function, in the proposed language for the second prong of the diagnosis.  The proposed criteria require:

 [C]oncurrent deficits in at least two domains of adaptive functioning of at least two or more standard deviations, which generally translates into performance in the lowest 3 % of a person’s age and cultural group, or standard scores of 70 or below.  This should be measured with individualized, standardized, culturally appropriate, psychometrically sound measures.  Adaptive behavior domains typically include:
  • Conceptual skills (communication, language, time, money, academic)
  • Social skills (interpersonal skills, social responsibility, recreation, friendships)
  • Practical skills (daily living skills, work, travel).
The ambiguity arises when one considers what is meant by a “domain.”  DSM-IV required deficits in at least two “areas” (similar to what appear to be sub-domains above).  In contrast, AAIDD has rightfully relied upon, based on factor analytic studies and the work of Stephen Greenspan, deficits in the domains of Conceptual, Social and Practical skills.  However, the AAIDD requires adaptive function deficits in only one as opposed to two domains.  The current AAIDD manual operationally defines significant limitations in adaptive behavior as “performance that is approximately two standard deviations below the mean of either (a) one of the following three types of adaptive behavior: conceptual, social, or practical or (b) an overall score on a standardized measure of conceptual, social, or practical skills.  The assessment instrument’s standard error of measurement must be considered when interpreting the individual’s obtained scores” p. 43.  The impact of requiring deficits in two of the three domains, as apparently required by the proposed criteria, has long been recognized.  For example, The 2002 AAMR Mental Retardation: Definition, Classification, and Systems of Supports manual described the impact that requiring deficits in two or more of the three domains would have on the prevalence of intellectual disabilities:

…the probability of a person having significant deficits (2 SDs below the mean) in two or in all three domains of adaptive behavior is extremely low compared to the probability of scoring two standard deviations or below on only one domain.  In fact, simulation studies have demonstrated that the probability of a person scoring two standard deviations below the mean on more than one domain would be so low that almost no one with an IQ in the upper mental retardation range would be identified as having mental retardation (K.F. Widaman, personal communication, November 9, 2001) (p. 78).

The proposed DSM-5 language, rather than having “Consistency with AAIDD practices,” as stated in the rationale, appears to fly in the face of those standards and would insure that “almost no one with an IQ in the upper mental retardation range would be identified as having” an intellectual disability.  In addition, by failing to note the importance of considering the “instrument’s standard error of measurement” the proposed criteria once again suggests a “bright-line” cut-off for both the intellectual and adaptive functioning requirements.  Effectively, and contrary to the stated rationale, these proposals represent changes in the cutoff scores.  In an attempt to establish unambiguous criteria for the intellectual and adaptive behavior diagnostic prongs the proposed language ignores a fundamental understanding of the nature of test scores, i.e., that some degree of imprecision is inherent.

The practical impact of these proposed criteria within a clinical context would be to reduce the number of individuals diagnosed with an Intellectual Disability.  Within an Atkins context, these changes would make more individuals eligible for the death penalty.

The Work Groups for DSM-5 are soliciting comments upon the proposed diagnostic criteria until April 20, 1010 at www.dsm5.org.  I would urge anyone with concerns regarding the criteria to submit their comments.

Dale G. Watson, Ph.D.
Clinical and Forensic Neuropsychologist
watson.dale@comcast.net

Technorati Tags: , , , , , , , , , , , , , , , , , , , , , , , ,