Wednesday, January 14, 2009

Premorbid prediction of WISC-IV IQs in children: Be careful


The post-hoc, retrospective prediction of a person's global IQ score, after significant brain injury (TBI), has been an ongoing area of study and discussion in the adult neuropsychology literature for many years. A new small scale study (n=40 TBI and 40 controls; thus, significant caution is thus urged) in Psychological Assessment reports an attemp to predict "premorid IQ" in children on the WISC-IV.

The abstract for the article by Schoenberg et al. (2008) is below. The bottom line message appears to be caution in attempts to predict a child's intelligence prior to TBI after TBI has occurred. Statistical formula's are available, but do not always provide decent estimates. The authors urge appropriate caution and the need to develop premorbid estimates that include more than just post-injury WISC-IV scores and select subject demographics (plugged into equations) ---e.g., consideration of prior group achievement tests scores; school grades; etc.

As per usual, the authors make a call for further research....which appears appropriate given the small sample sizes and the accuracy of the WISC-IV equation based prediction methods.
  • Abstract: Determination of neuropsychological impairment involves contrasting obtained performances with a comparison standard, which is often an estimate of premorbid IQ. M. R. Schoenberg, R. T. Lange, T. A. Brickell, and D. H. Saklofske (2007) proposed the Child Premorbid Intelligence Estimate (CPIE) to predict premorbid Full Scale IQ (FSIQ) using the Wechsler Intelligence Scale for Children—4th Edition (WISC–IV; Wechsler, 2003). The CPIE includes 12 algorithms to predict FSIQ, 1 using demographic variables and 11 algorithms combining WISC–IV subtest raw scores with demographic variables. The CPIE was applied to a sample of children with acquired traumatic brain injury (TBI sample; n  40) and a healthy demographically matched sample (n  40). Pairedsamples t tests found estimated premorbid FSIQ differed from obtained FSIQ when applied to the TBI sample (ps.01). When applied to healthy peers, estimated and obtained FSIQ did not differ (ps.02). The demographic only algorithm performed well at a group level, but estimates were restricted in range. Algorithms combining single subtest scores with demographics performed adequately. Results support the clinical application of the CPIE algorithms. However, limitations to estimating individual premorbid ability, including statistical and developmental factors, must be considered.
Technorati Tags: , , , , , , , , ,