The MMPI and MMPI-2 are among the most commonly used personality inventories in the world today (Lezak, 1995). Despite the importance of issues relating to psychosocial adjustment to neuropsychological assessment, relatively little research has been conducted in the applicability of these inventories to the evaluation of traumatic brain injury (TBI). This is particularly alarming when one considers the frequency with which the MMPI/MMPI-2 is employed in assessing TBI in the medicolegal arena.

Much of the MMPI/MMPI-2 research conducted with brain-injured subjects has considered the issues of malingering and the correction of profiles, with little emphasis given to the forensic utility of this instrument. A few researchers have attempted to examine the MMPI/MMPI-2 and brain-injured populations (Gass, 1992; Gass & Russell, 1991; Hamilton, Finlayson & Alfano, 1995; Watson & Plemel, 1978). These studies have tended to generate critical item sets or methods of profile correction which, at least in Australia, have found little favour with clinicians. No study has attempted to explicitly develop a scale sensitive to the psychosocial self-reports of TBI cases who are currently engaged in litigation.

The goal of the current study was to develop a MMPI-2 scale relevant to the assessment of TBI cases evaluated in the medicolegal setting. This was achieved in two stages: first, the construction of an MMPI-2 scale utilising a forensic TBI sample, and; second, an examination of the utility of the scale in assessing its specificity through comparisons with other common forensic groups such as chronic pain (CP), and post-traumatic stress disorder (PTSD). In this second stage other brain-injury-related MMPI-2 scales that are available to the clinician are evaluated

The first study utilised the MMPI-2 protocols of 100 head-injured patients, who were evaluated in a forensic setting between 1993 and 1996. All clients who met the criteria for CNS Trauma Report (1985) classification were included. Only protocols where VRIN or TRIN had T-scores greater than 80 were excluded. A comparison group of 88 university students was also utilised to generate preliminary normative data. Descriptive statistics for both samples can be found below (table 1).

1) Using the TBI sample, the frequency of "True" and "False" responses to all 567 MMPI-2 items were derived separately for males and females.

2) These percentages were then compared with the contemporary normative data in the MMPI-2 Manual (Appendix I), to produce an index of discrimination (IOD).

3) Using the IOD distribution, items at less than the 5th percentile and greater than the 95th percentile were selected for inclusion in the scale. Of these 73 items, 39, to which both males and females responded alike, were kept. Thus, ensuring a reduction in the item pool and maximal discrimination between TBI and the MMPI-2 norms.

4) To improve the reliability of the scale, a further 3 items were removed - leaving 36 items in total for the scale.

5) Exploratory factor analysis provided 3 factors with eigenvalues greater than 1.0. Examination of individual items revealed consistent content themes that related to cognitive, emotional and physical issues, and these were used to develop the three subscales. Based on item content 4 items were reallocated to another factor resulting in an increase in subscale reliability.

The 36 items comprising the Traumatic Brain Injury - Forensic Scale, its subscales, and their scored direction are provided below along with measures of internal consistency. Scale/subscale intercorrelations for normals and TBI patients are also provided (Table 2). Table 3 illustrates the degree of overlap between the current TBI(f) scale item content and those of other measures available in the literature. Table 4 shows the correlations between TBI(f) scale and subscales and other MMPI-2 validity, basic, content, and supplementary scales.
 

TBI(f) Scale 36 items; 19 false & 17 true alpha coefficient: TBI = .85, Normals = 0.83

Cognitive Subscale 13 items alpha coefficient: TBI = .82, Normals = 0.77 11 True: Items 31, 140, 147, 168, 299, 308, 325, 472, 475, 533, 565 2 False: Items 43, 165

Emotional Subscale 8 items alpha coefficient: TBI = .74, Normals = 0.67 5 True: Items 130, 305, 377, 469, 513 3 False: Items 9, 148, 388

Physical Subscale 15 items alpha coefficient: TBI = .75, Normals = .63 2 True: Items 53, 247 13 False: Items 10, 33, 45, 57, 141, 152, 173, 176, 177, 179, 224, 295, 561

428 participants were included in the validation study which formed four separate groups: normal university students (N = 200); TBI cases (N = 100); PTSD cases (N = 25); and CP cases (N = 103). All clinical cases were derived from a forensic practice in Brisbane, Australia and had confirmed diagnoses. The characteristics of each of these samples are listed in Table 5. The data analysed in this study were the full MMPI-2 protocols completed by all subjects and all scales, both standardised and experimental were computed. Only protocols with TRIN or VRIN greater than 80 were excluded from the study. Although the normal group of university students represent an opportunistic sample, it should be noted that age and education are not variables for which differing normative data has been generated for the MMPI-2. For the purposes of this study it is sufficient that they are a "normal" group with no reported neurologic or psychiatric histories.

In addition to examining the TBI(f) scale and subscales, the Cerebral Impairment (CI) Index (Jacobucci, 1993) and the Gass (1991) Neurological (Neur) and Psychiatric (Psych) items were also computed and compared.

Sufficient data to determine the severity of injury was only available for subjects in the TBI group, but as no significant differences were found on this variable, all data presented will be for this group as a whole. However, the average interval from the time of injury to the time of testing was 29.05 months (SD=30.66), and injuries ranged in severity from mild (N = 45) to moderate (N = 29) and severe (N = 26).

Significant differences were found between all MMPI-2 brain injury measures (TBI(f), CI, Gass items) for TBI patients and Normals but no significant differences were found between the three clinical groups on any of these measures. Only the TBI(f) subscales were able to differentiate between the three clinical groups. For TBI(f) - 1, the cognitive subscale a significant difference was found between the TBI group and CP groups (t (201) = -4.09, p<.001) reflecting greater reports of cognitive difficulties following traumatic brain injury than with chronic pain. No significant difference was found on this subscale for TBI and PTSD. For TBI(f) -2, the emotional subscale, a significant difference was found between TBI and PTSD (t (123) = 2.27, p = .025) indicating that the PTSD group reported greater emotional distress than did the TBI group. No between group differences were found for the TBI(f) - 3 Physical subscale indicating all three groups are reporting equivalent levels of physical concerns and symptoms. The means and standard deviations for all brain injury measures and groups are displayed in Table 6. The test operating characteristics in Table 7 demonstrate that the TBI(f) scale and subscales are effective in discriminating between unimpaired individuals and those who have sustained a traumatic brain injury. Only the Gass Neurological items consistently match the TBI(f) for overall predictive power. The scale performs less well in differentiating between clinical groups but as indicated above a differential pattern on the TBI(f) subscales may be the key to discriminating group membership.

Table 7. Evaluation of Test Operating Characteristics of the MMPI-2 Brain Injury Measures Between Both Normal and Clinical Groups

Key: Sens. - sensitivity; Spec. - specificity; PPP - positive predictive power; NPP - negative predictive power; OPP - overall predictive power

The development of a TBI(f) scale for the MMPI-2 provides the clinician with a method of evaluating the degree to which a particular client is experiencing psychosocial difficulties commonly reported by TBI patients in a forensic setting. The scale and its subscales have psychometric properties consistent with those of other MMPI-2 scales and compare well with others' attempts to identify items that relate to TBI.

In other studies examining MMPI-2 performance in TBI patients, researchers have often excluded those with F scale T scores greater than 70. In the first study, such a criterion would have excluded a third of the sample calling into question external validity of the study. The correlations between the TBI(f) scale and validity scales while significant were not substantial., accounting for no more than 20% of their variance. How this relates to the appropriateness of standard validity indices in this sample is uncertain and requires further study with other TBI samples in which financial compensation is not involved as well as compensation non-TBI cases which was the focus of Study 2.

Through evaluation of test operating characteristics in study 2, it is clear that the TBI(f) scale can effectively discriminate between the MMPI-2 scores of TBI patients and normals with an overall predictive power of .87. It is also clear that the scale performs less well when discriminating between TBI and other clinical groups commonly encountered in forensic practice. Many of the problems reported by TBI patients are not unique to this disorder and are commonly found in other clinical groups such as CP and PTSD. Nonetheless, the TBI(f) subscales demonstrate the potential to discriminate between the different clinical groups evaluated in this study and represent a major advantage over the Gass Neurological Items. Clinicians are advised that at present the TBI(f) scale is recommended for forensic assessment of the degree of cognitive, emotional, and physical distress reported by individuals following a traumatic brain injury. The use of the scale to determine whether or not a particular individual's MMPI-2 profile is more consistent with TBI than PTSD, or Chronic Pain is not recommended and will be the focus of future research.

Becker, D. P., & Povlishock, J. T. (Eds.). (1985). Central nervous system trauma status report. National Institute Of Neurological and Communicative Disorders and Stroke. National Institutes of Health.

Gass, C. S. (1991). MMPI-2 interpretation and closed head injury: A correction factor. Psychological Assessment, 3, 27-31.

Gass, C. S. (1992). MMPI-2 interpretation of patients with cerebrovascular disease: A correction factor. Archives of Clinical Neuropsychology, 7,17-27.

Gass, C. S., & Russell, E. W. (1991). MMPI profiles of closed head trauma patients: Impact of neurologic complaints.Journal of Clinical Psychology, 47,253-260.

Hamilton, J. M., Finlayson, A. J.,& Alfano, D. P. (1995). Dimensions of neurobehavioural dysfunction: Cross-validation using a head-injured sample. Brain Injury,, 9,479-485.

Jacobucci, G. (1993). Cerebral impairment: A new content scale for the MMPI-2. Masters Abstracts International, 33, 674-792.

Lezak, M. D. (1995). Neuropsychological assessment (3rd ed). New York: Oxford University Press.

Watson, C. G., & Plemel, D. (1978). An MMPI scale to separate brain-damaged from functional psychiatric patients in neuropsychiatric settings. Journal of Consulting and Clinical Psychology, 46,1127-1.