Detection of malingering in a Spanish population using three specific malingering tests

Archives of Clinical Neuropsychology 22 (2007) 379–388

Detection of malingering in a Spanish population usingthree specific malingering tests

Raquel Vilar-Lopez a,b,∗, S. Santiago-Ramajo a, Manuel Gomez-Rıo a,Antonio Verdejo-Garcıa b,c, Jose M. Llamas a, Miguel Perez-Garcıa b,c,∗

a Servicio de Medicina Nuclear, Hospital Universitario “Virgen de las Nieves”, Granada, Spainb Departamento de Personalidad, Evaluacion y Tratamiento Psicologico,Universidad de Granada, Granada, Spain

c Neurosciences Institute. Universidad de Granada, Granada, Spain

Abstract

The detection of feigned cognitive impairment remains difficult and may be even more challenging in certain population groups.Studies on the use of neuropsychological tests in ethnic groups for which they were not designed have shown variations in performanceassociated with cultural differences. With this background, our group studied a Spanish population by applying commonly usedprocedures [Victoria Symptom Validity Test (VSVT), Test of Memory Malingering (TOMM), and the b test] in a group with post-concussion syndrome (PCS) (whether litigants or not) and in a group of analog malingerers (AN). These tests appeared to functionadequately in this Spanish population, who showed similar performances to results published for North Americans.© 2007 National Academy of Neuropsychology. Published by Elsevier Ltd. All rights reserved.

Keywords: Malingering; Post-concussion syndrome; Spanish speakers; VSVT; TOMM; b test

1. Introduction

Forensic neuropsychology has experienced enormous growth over recent years. Thus, the percentage of articles onneuropsychology in the most widely read neuropsychological journals (Archives of Clinical Neuropsychology, Journalof Clinical and Experimental Neuropsychology, and The Clinical Neuropsychology) increased from 4% in 1990 to14% in 2000. The predominant issue in these studies is malingering, addressed by 86% of the forensic papers (Sweet,King, Malina, Bergman, & Simmons, 2002).

The upsurge of interest in this issue results from the increasingly frequent demand for neuropsychologists to giveexpert opinion in personal injury litigation, especially in cases of mild or moderate brain damage. Thus, a largeproportion of individuals undergoing neuropsychological assessment in the United States are implicated in legal casesthat would afford them considerable economic benefits if cognitive damage were demonstrated, whether genuine ornot. In many cases, neuropsychological test data are the only objective source of evidence of deficits, especially incases of mild brain damage, where neuroimaging techniques generally give negative results and neurological signs areusually absent.

Within the spectrum of malingering, individuals can opt to feign cognitive difficulties, psychiatric complications, orsomatic problems, such as chronic pain. Unfortunately, detection of the last two types is poorly developed and remains

∗ Corresponding authors at: Departamento de Personalidad, Evaluacion y Tratamiento Psicologico, Facultad de Psicologıa de la Universidad deGranada, Campus de Cartuja s/n, 18071 Granada, Spain. Tel.: +34 958 242 948; fax: +34 958 243 749.

E-mail address: [email protected] (R. Vilar-Lopez).

0887-6177/$ – see front matter © 2007 National Academy of Neuropsychology. Published by Elsevier Ltd. All rights reserved.doi:10.1016/j.acn.2007.01.012

mailto:[email protected]

dx.doi.org/10.1016/j.acn.2007.01.012

380 R. Vilar-Lopez et al. / Archives of Clinical Neuropsychology 22 (2007) 379–388

to be fully addressed by neuropsychologists, who have largely focused on detecting the feigning of cognitive deficits,which has been widely and extensively researched since the 1980s. This intensity of research efforts is related to the veryhigh social and personal costs of inaccurate testing (false positives or false negatives). Thus, in the state of Louisianaalone, malingering has been estimated to cost $107.2 million per year when sentences and verdicts, expenditure onunnecessary medical treatments, and defense costs are included (Gouvier, Lees-Haley, & Hammer, 2003). If this figureis extrapolated to the entire nation, the costs of malingering in the United States may be as high as $5.36 billion peryear, without taking account of working days lost or the cost of substituting for malingerers (Gouvier et al., 2003). Ifthis can be considered the cost of false negatives, false positives incur important costs that are more social or personal,such as the failure to treat patients incorrectly identified as malingerers and their negative labeling.

The difficulty of detecting malingering may be even greater in specific populations, such as Hispanics. Studies onthe use of neuropsychological tests with other ethnic groups than the tests were designed for have shown differencesin performance associated with cultural differences (Puente & Perez-Garcia, 2000a). These differences may lead todiagnostic errors, and it has been recommended that specific scales and cutoff points be used for each ethnic group(Puente & Perez-Garcia, 2000b). Thus, the NAN Policy & Planning Committee (Bush et al., 2005) stated that the use ofSymptom Validity Tests with culturally diverse populations for whom validation data do not exist must be carefully con-sidered (p. 425). In the case of Hispanic populations, this difficulty is compounded by the low development of forensicneuropsychology in Spanish-speaking countries (Verdejo-Garcıa, Alcazar, Gomez-Jarabo, & Perez-Garcıa, 2004).

With the above background, we decided to investigate the performances of a Spanish population in the commonlyused Victoria Symptom Validity Test (VSVT), Test of Memory Malingering (TOMM), and b test, including patientswith post-concussion syndrome (PCS) involved in litigation, patients with PCS not involved in litigation, and a groupof analog malingers.

2. Method

2.1. Subjects

Three groups were studied: a group of patients with PCS not involved in any PCS-related litigation (PCS-NLI); agroup of patients with PCS and involved in PCS-related litigation (PCS-LI), and a group of analog malingerers (AN).

The PCS-NLI group comprised 12 individuals (6 males and 6 females) with mean (±standard deviation) age of31.16 ± 10.67 years and mean 9.58 ± 3.62 years of schooling. Inclusion criteria were diagnosis of mild traumatic braininjury according to the American Congress of Rehabilitation Medicine (Kay et al., 1993) (loss of consciousness for≤30 min, Glasgow Coma Scale of 13–15 points, and post-traumatic amnesia for <24 h), diagnosis of PCS according toICD-10 (World Health Organization, 1992), and age between 18 and 55 years. Exclusion criteria were pregnancy (dueto radiology protection criteria), head CT findings or presence of psychiatric disorders, history of substance abuse, andengagement in PCS-related litigation. The PCS-LI group comprised 14 individuals (3 female and 11 male) with meanage of 30.35 ± 9.77 years and mean 8.92 ± 2.39 years of schooling. The inclusion and exclusion criteria were the sameas for the PCS-NLI group except that involvement in PCS-related litigation was an inclusion criterion. The AN groupcomprised 35 university students (3 male and 32 female) with mean age of 23.22 ± 1.76 years and mean 16.86 ± 0.41years of schooling. Inclusion criteria were to be a 4th-year psychology student and therefore possess knowledge ofbrain damage and neuropsychology; and age between 18 and 55 years. Exclusion criteria were the same as for theother groups, except that pregnancy was not an exclusion criterion.

Differences were found between PCS groups and the AN group on age [F(2,58) = 12.30; p < .000], and education[F(2,58) = 115.01; p < .000], being the AN group younger and better educated than the other groups. Nevertheless,these variables are not related to performance on the VSVT, TOMM, and b Test (Boone, Lu, & Hezberg, 2002; Slick,Hopp, Strauss, & Thompson, 1997; Tombaugh, 1996). Moreover, the AN group is younger and better educated thanPCS groups, so it is difficult to explain their worse performance on the malingering tests attending to demographicvariables.

2.2. Instruments

The VSVT (Slick et al., 1997) is a computer-administered and scored test. It includes a total of 48 items, pre-sented in three blocks of 16 items each. During the study trial, a single five-digit study number is presented for 5 s

R. Vilar-Lopez et al. / Archives of Clinical Neuropsychology 22 (2007) 379–388 381

in the center of the computer screen. This presentation is followed by the retention interval and then by the recog-nition trial, in which the target and a five-digit distractor are displayed. The respondents are asked to choose thenumber they saw in the study trial. The retention interval is 5 s in the first block, 10 s in the second, and 15 s in thethird.

The VSVT’s reliability and validity were found to be excellent in some studies (for example, 91 and 98%,respectively, in Inman, Vickery, Lamb, Edwards, & Smith, 1998).

The TOMM (Tombaugh, 1996) is an instrument designed to provide an assessment whether an individual is falsifyingsymptoms of memory impairment. It consists of 50 pictures of common objects that the subject must remember. Afterthe presentation of the 50 pictures, the individual has to recognize which one is the correct picture between twoalternatives. The TOMM consists of two learning trials and an optional retention trial and has good face validity as atest of learning and memory.

Finally, the b Test (Boone et al., 2002) is a measure to identify malingering requiring recognition of over-learnedinformation. This test is a letter recognition and discrimination task that consists of a 15-page Stimulus Booklet. Theexaminee is asked to circle all of the “b” that appear on each page, working as quickly as possible.

2.3. Procedure

The assessment of the PCS-NLI and PCS-LI groups comprised a functional neuroimaging study (SPECT) (datanot analyzed in the present study) and, on a different day, a neuropsychological battery of tests. On both occasions,the patient received a clear explanation of the evaluation procedure and written consent was obtained. This evaluationwas conducted, at least, 6 months post-injury, so we can consider that results do not show residual deficits fromuncomplicated mild traumatic brain injury.

At their neuropsychological assessment session, individuals in both PCS groups first underwent an interview togather the following: demographic data, medical disorders and treatments received, current treatments, subjective neu-ropsychological complaints, brain injury data (consciousness level, neurological deficit, general medical history), andPCS (sociological setting, social impact and administration of Sheffield Brain Injury Questionnaire [Rivermead ques-tionnaire]). Subjects also underwent an extensive neuropsychological assessment that included the VSVT, TOMM, andb test, specific malignancy tests. The interview and neuropsychological battery lasted approximately 2.5 h; therefore, arest period of 20 min was provided half way through the test. On some occasions, as required, more rest periods wereallowed or two sessions were used for the battery.

Following the recommendations of Rogers (1997), individuals in the AN group were first given instructions thatdescribed the situation and informed them about PCS. Course credits awarded to them for their collaboration werenotably increased if their feigning was successful, providing a real incentive. Participants had as long as they considerednecessary to read the instructions before undergoing the malingering tests. After completion of the assessment, theycompleted a questionnaire to test their understanding of the instructions and the degree of their involvement during thetests. This process lasted approximately 1.25 h.

2.4. Variables

The variable selected for the VSVT was the total score of correct responses. The cutoff point was 90%, with a lowerscore indicating malingering. Although the test manual (Slick et al., 1997) recommends the use of binomial probability,as the reliability and validity are not subject to the type of sample used, numerous researchers have reported that thisstrategy gives rise to an unacceptably high number of false negatives (Bender & Rogers, 2004; Gervais, Rohling, Green,& Ford, 2004; Greiffenstein, Baker, & Gola, 1994; Guilmette, Hart, & Giuliano, 1993; Guilmette, Hart, Giuliano, &Leininger, 1994; Hiscock, Branham, & Hiscock, 1994; Holmquist & Wanlass, 2002; Martin, Bolter, Todd, & Gouvier,1993; Martin, Hayes, & Gouvier, 1996; Rose, Hall, & Szalda-Petree, 1995; Slick, Hopp, Strauss, Hunter, & Pinch,1994). Another option offered by the manual is to consider a cutoff score of 44 (90%) correct responses, below whichmalingering is indicated. According to the manual, a score of 44 is in the first percentile, indicating that 99% ofparticipants in the reference group have higher scores. It concludes that litigants with mild traumatic brain injury whodo not achieve these scores can be suspected of exaggerating their symptoms. The same criterion was followed in thepresent study.


Table 1Means, standard deviations, ANOVAS, and post-hoc analysis of the malingering tests

Test PCS-NLI [mean (SD)] PCS-LI [mean (SD)] AN [mean (SD)] F p Scheffe

VSVT 45.88 (1.83) 42.3 (5.53) 30.77 (7.38) 26.22 .000 (1 = 2) > 3TOMM 49.91 (0.28) 45.91 (4.88) 34.4 (8.57) 27.3 .000 (1 = 2) > 3The b test 52.65 (21.70) 87.16 (58.03) 393.17 (318.99) 11.95 .000 (1 = 2) < 3

Note. PCS-NLI, patients with post-concussion syndrome (PCS) not in litigation; PCS-LI, patients with PCS in litigation; AN, analog malingerers;NS, not significant; SD, standard deviation; VSVT, Victoria Symptom Validity Test; TOMM, Test of Memory Malingering.

The variable selected for the b test (Boone et al., 2002) was the e-score, recommended by the manual as the mainscore for test interpretation, calculated as follows:

e-score = [(total errors d + total errors of commission) × 10]

+ total errors of omission + mean time for each stimulus

The cutoff point indicating malingering was 90, specified by the test manual for brain injury patients.The variable for the TOMM test (Tombaugh, 1996) was the total number of correct responses in Trial 2 of the test.

The cutoff point was 45, with a lower score indicating malingering, as recommended in the test manual.

3. Results

3.1. Differentiation of malingering variables among study groups

Differences among the groups in the malingering tests were tested using three univariate ANOVAS, with the groupas independent variable (PCS-NLI vs. PCS-LI vs. AN) and the performance in malingering tests as dependent variables,i.e., total number of correct responses in the VSVT, total number of correct responses in Trial 2 of the TOMM, ande-score in the b test. Statistically significant differences among the groups were found in VSVT results [F(2.51) = 26.22;p < .001], and post-hoc analysis (Scheffe test) showed the mean of the AN group to be significantly lower than the meansof the other two groups (PCS-NLI and PCS-LI), which showed no differences between them. Significant differenceswere also found in the TOMM Trial 2 results [F(2.56) = 27.30; p < .001], and the Scheffe test showed the mean of theAN group to be significantly lower than the means of the other two groups, which showed no differences betweenthem. Finally, significant differences were found in the e-score [F(2.56) = 11.95; p < .001], and the Scheffe test showedthe mean of the AN group to be significantly higher than the means of the other two groups, which again showed nodifferences between them (Table 1). Nevertheless, the lack of differences between the PCS groups could be due, in ouropinion, to the small sample sizes.

3.2. Relationship between malingering test classification and study groups

In order to detect any association between the group variable and the classification as malingerer, three contingencyanalyses were performed crossing the group variable (PCS-NLI, PCS-LI, or AN) with the classification of malingereror non-malingerer obtained in each test. Statistically significant associations were found between the groups and theclassifications obtained. Thus, in the VSVT, 97.1% of the AN group and 50% of the PCS-LI group were classified asmalingerers, whereas 100% of the SPC-NLI group were classified as non-malingerers [χ2(2) = 36.69; p < .000]; in theTOMM, 88.6% of AN were classified as malingerers, 100% of PCS-NLI as non-malingerers, and 58.3% of PCS-LIas non-malingerers [χ2(2) = 31.84; p < .000]; and in the b test, 80% of AN were classified as malingerers and 91.7 and66.7% of the PCS-NLI and PCS-LI groups as non-malingerers, respectively [χ2(2) = 21.74; p < .000] (Table 2).

3.3. PCS-LI group

Because some members of the PCS-LI group were classified as non-malingerers and others as malingerers, thecomposition of this group was examined. One subgroup was formed of those classified as malingerers in two out ofthe three tests, considered to indicate a high probability of malingering (PCS-LI-M) (n = 7). This multiple criteria as


Table 2Percentages of individuals classified as malingerers and non-malingerers by the tests within each study group

Test SPC-NLI SPC-LI AN χ2 p

Maling Not mal Maling Not mal Maling Not mal

VSVT 0 100 50 50 97.1 2.9 36.69 .0001TOMM 0 100 41 58 86 11.4 31.84 .0001The b test 8.3 91.7 33.3 66.7 80 20 21.74 .0001

Note. PCS-NLI, patients with post-concussion syndrome (PCS) not in litigation; PCS-LI, patients with PCS in litigation; AN, analog malingerers;Maling, percentage of individuals who were malingering according to the different tests; Not mal, percentage of individuals who were not malingeringaccording to the different tests; VSVT, Victoria Symptom Validity Test; TOMM, Test of Memory Malingering.

indicative of malingering is in agreement, for example, with study by Slick, Sherman, and Iverson (1999) or withLarrabee (2003). Another subgroup was formed with those who did not fail any of the three malingering tests and weretherefore considered to have a low probability of malingering (PCS-LI-NM) (n = 7). Three nonparametric analysesfor two independent samples (Mann–Whitney U-test) were then performed, with the group variable (PCS-LI-M vs.PCS-LI-NM) as independent variable and the malingering test scores as dependent variables. Statistically significantdifferences were found for the VSVT [Z = −2.581; p = .006], TOMM [Z = −3,205; p = , 001], and b test [Z = −3,130;p = .001] scores (Table 3).

The scores of the high-probability subgroup (PCS-LI-M) were then compared with those of the AN group, performingthree nonparametric analyses for two independent samples (Mann–Whitney U-test), with the group variable (PCS-LI-M vs. AN) as independent variable and malingering test scores as dependent variables. No statistically differentdifferences were found between these groups (Table 3).

Finally, scores of the low-probability subgroup (PCS-LI-NM) were compared with those of the non-litigant group(PCS-NLI), performing three nonparametric analyses for two independent samples (Mann–Whitney U-test), with thegroup variable (PCS-LI-NM vs. PCS-NLI) as independent variable and the malingering test scores as dependentvariables. No significant differences were found between these groups (Table 3).

According to the above data, the probability of feigning in the litigating PCS group in the present sample was 50%(7 out of 14 patients).

3.4. Comparison of malingering test results between Spanish and North American samples

New mean contrast analyses were performed to test whether there were differences between the means obtainedfor our Spanish population and those of the North American populations published in the manuals of the three testsused. In the VSCT and b test, no statistically significant differences were found between the mean scores obtained bythe Spanish and North American samples for any of the groups, and in the TOMM, only the scores of the PCS-LIMgroup significantly differed (p = .0025) from those of the equivalent North American group (see Table 4). In order totest whether this significant difference was due to the heterogeneity of the PCS-LI group, two mean contrast analyseswere performed, comparing the mean obtained by Spanish individuals scoring as malingerers in at least two out ofthree of the tests with the North American sample. In this test, the statistical significance of the difference in TOMMscores between the North American and Spanish sample disappeared (p = .4147) (Table 5).

Table 3Nonparametric test results between the groups

PCS-LI-NM vs. PCS-LI-M PCS-NLI vs. PCS-LI-NM PCS-LI-M vs. AN

Z p Z p Z p

VSVT −2.581 .006 −1.740 .106 −1.350 .182TOMM −3.205 .001 −1.902 .340 −1.045 .312e-score −3.130 .001 −0.761 .482 −1.267 .216

Note. PCS-LI-NM, patients with post-concussion syndrome (PCS) in litigation who have not scored as malingerers in ≥2 tests; PCS-LI-M, patientswith PCS in litigation who have scored as malingerers in ≥2 tests; PCS-NLI, patients with PCS not in litigation; AN, analog malingerers; VSVT,Victoria Symptom Validity Test; TOMM, Test of Memory Malingering.


Table 4Means and standard deviations of the selected scores in the three malingering tests for the study groups in the Spanish and North American samples

Test GR.1 GR.2 GR.3

Spanisha N-Am. p Spanisha N-Am. p Spanisha N-Am. p

VSVT 45.88 (1.83) 46.16b (2.59) .7454 42.3 (5.53) 43.50c (6.08) .4734 30.77 (7.38) 31.26d (9.05) .7971TOMM 49.91 (0.28) 49.5e (1.1) .2197 45.91 (4.88) 32.8f (13.4) .0025** 34.4 (8.57) 35.3d (9.4) .6957Test b 52.65 (21.70) 58.5e (18.7) .4260 87.16 (58.03) 612.5g (1069.6) .0702 393.17 (318.99) No data

Note. GR.1, group with brain damage not suspected of malingering; GR.2, group with brain damage and high probability of malingering; GR.3,analog malingerer group; N-Am., North American sample.

a Spanish subjects all had uncomplicated mild traumatic brain injury, but North American groups include patients with significant neurologicaldysfunction. The lack of differences between this groups could be showing that performance in SVT do not depend on severity of injury.

b Patients with seizure disorder and patients with head injury or other brain injury dysfunction.c Patients seen for medicolegal neuropsychological evaluations following possible head injury.d Analog undergraduate students.e Non-litigating traumatic brain injury patients.f Patients at risk of malingering defined by the presence of one or more of the following factors: pursuit of a personal injury suit; pursuit of a

disability petition; symptoms discrepant from known neurologic disease.g Participants with at least two test results indicating non-credible cognitive symptoms, drawn from at least two tests.

Table 5Means and standard deviations of the selected scores in the different tests for the corrected GR.2 group (with two tests indicating malingering) inthe Spanish and North American samples

Test Spanish North American p

TOMM 37.57 (8.42) 32.8 (13.4) .4147Test b 247.76 (196.57) 612.5 (1069.6) .3717

4. Discussion

This study provides the first demonstration of the efficacy of three commonly used malingering tests in a Spanishseries comprising three groups: individuals with PCS involved in related litigation, individuals with PCS not involvedin related litigation, and AN. Moreover, the scores obtained are compared with those published in the original testmanuals for the North American populations on which they were based. Results of the VSVT, TOMM, and b tests inthe Spanish population differed by subgroup, with a worse performance by individuals with PCS involved in litigationcompared with those who were not, and with the AN showing significantly worse scores compared with the PCSgroups. Furthermore, statistically significant associations were found between these groups and the classification ofmalingerer versus non-malingerer in all three tests. It was also observed that the fact of being a litigant was not sufficientto be classified as having a high probability of being a malingerer. The only significant differences found between thedata of the Spanish population and those provided by the test designers (based on North American groups) were in theTOMM, for the group considered to have a high probability of malingering.

4.1. Efficacy of malingering tests in the Spanish population

In all three malingering tests, statistically significant differences were found in the results of the groups, the mainfinding of the study. On the contrary, no significant differences were found between litigants and non-litigants withPCS, whereas the scores of both groups significantly differed from those of the AN. A possible explanation for thisdifference, also observed by Slick, Hopp, Strauss, and Spellacy (1996) with VSVT), is that analog studies lack externalvalidity. Nevertheless, this study followed all recommendations proposed by the most widely accepted authorities toincrease the rigor of the research and maximize its external validity (Rogers, 1997). Also, the lack of differences couldbe due to the small sample size. However, the effect sizes was δ = 0.97 for the VSVT, δ = 1.55 for the TOMM, andδ = − 0.86 for the b test, which are a large effect size for the VSVT and b test and a very large effect size for the TOMMin accord with Bezeau and Graves (2001) recommendations.


Moreover, a more detailed analysis suggested that the results pointed in the expected direction: the scores of thenon-litigants were close to the maximum in the tests, indicating an optimal effort during the assessment, whereasthe litigants obtained scores that were close to the cutoff point for the classification (Table 1). Thus, the mean totalscore for litigants was 45.88 for non-litigants, 42.3 for litigants, and 30.77 for the AN. Given that the maximumscore was 48, the score of non-litigants was close to its ceiling, whereas the litigants had lower scores. If a cutoffscore of >43 (i.e., 90% correct responses) was considered to indicate malingering, the mean of the litigants wouldbe below this point and the mean of the AN would be considerably above it. This trend was also observed in theTOMM, with the non-litigants obtaining the highest mean, (49.91/50 points), the litigants showing a mean (45.91)close to the selected cutoff point of 45, and the AN scoring well below this cutoff point (34.4). The same resultswere observed for the b tests, in which malingering was indicated by an e-score above 90, with a score of 52.65 fornon-litigants, a score close to the cutoff point for litigants (87.16), and a score considerably above this point for theAN (93.17).

The cutoff point considered for the VSVT was not the score preferred by its authors in their manual or other researchpapers (Slick et al., 1996, 1997). They indicated the following cutoff points for the interpretation of VSVT results:valid test = 48–30, questionable test = 29–18, and invalid test = 17–0 correct responses. If these indications had beenfollowed, the VSVT would have been considered valid for 100% of our non-litigants and 85% of the litigants, andonly 2.85% of the tests by AN would have been detected as invalid. Therefore, their recommended cutoff points givean unacceptable number of false positives. Our selection of a cutoff point of 90% correct responses for detectingmalingering in patients with mild brain injury appears to be much more appropriate, as the classification of AN andnon-litigants was practically ideal (no false positives and minimum false negatives). These results are in agreement withthose published by Greiffenstein et al. (1994), Grote et al. (2000), Guilmette et al. (1993, 1994), Inman et al. (1998),and Prigatano, Smason, Lamb, & Bortz (1997). Hence, the binomial probability proposed for the clinical interpretationof the test does not appear useful, as only a very small proportion of dissimulators scored significantly below chance(Binder & Rohling, 1996; Rogers, Harrell, & Liff, 1993; Slick et al., 1994), whereas the cutoff point of 90% correctresponses provides very high sensitivity and specificity (Greiffenstein et al., 1994; Grote et al., 2000; Guilmette et al.,1994; Inman et al., 1998; Prigatano et al., 1997).

The VSVT classified 97.1% of the AN and 50% of the litigant patients as malingerers, and 100% of non-litigantpatients as non-malingerers. In other words, the seven individuals suspected of malingering in the two groups of patientswith PCS all belonged to the litigant group, and no patient identified as a malingerer was in the non-litigant group.Moreover, the test classified almost all (97.1%) of the AN as malingerers, and only one individual escaped detection.The TOMM classified 86% of AN, none of the non-litigant patients, and 41% of litigant patients as malingerers. Thus,only 4 of the 35 AN escaped detection, whereas the 7 remaining individuals who scored below 45 points all belonged tothe litigant group and none to the non-litigant group. The b test classified 91.7% of non-litigants and 66.7% of litigantsas non-malingerers and detected that 80% of AN were malingerers. One of the individuals classified as a malingererby the b test belonged to the non-litigant group, although the score of this individual was very close to the cutoff point(score of 95 vs. cutoff point of 90), and no error of commission was committed. In this case, the high score was dueto the excessive time taken for the test, which may indicate an over-detailed approach rather than a feigning attempt.Moreover, this individual was not classified as a malingerer in the other tests.

Hence, the classification tests used in this study correctly classified AN as malingerers and non-litigant patients asnon-malingerers. However, the group of patients involved in litigation appears to be a heterogeneous group, in whichone subgroup is truly motivated to make the maximum effort in the tests, so that the scores reflect real damage, whereasthe other subgroup does not appear to be putting maximum effort into the test. This may be due to the fact that a highprobability of malingering within a group does not directly imply malingering, i.e., the litigant group is made up ofmalingerers and non-malingerers, making the scores of the group heterogeneous. This would explain the absence ofsignificant differences between the litigant and non-litigant groups as well as the classifications of malingerers andnon-malingerers in the litigant group. This hypothesis was tested by a deeper study of the litigant patient group, whichwas divided into two subgroups, one composed of patients scoring as malingerers in at least two of the tests (PCS-LI-M)and the other composed of the remaining litigant patients (PCS-LI-NM). Significant differences between these groupswere found in the scores of all three tests, and no differences in scores were found between the PCS-LI-NM group andthe non-litigant group or between the PCS-LI-M and the AN.

Therefore, it appears that the seeking of compensation is not sufficient to establish that a group can be suspected ofmalingering (Bianchini, Mathias, Greve, Houston, & Crouch, 2001; Greiffenstein et al., 1994; Prigatano et al., 1997;


Rogers, 1997), and patients should not be assigned to malingerer or control groups solely on the basis of their statusas litigants, which would be a failure in the internal validity of a study (Haines & Norris, 1995). Additional criteria orexternal markers of malingering must be adopted, such as those proposed by Greiffenstein et al. (1994), Prigatano et al.(1997), or Slick et al. (1999), in order to increase the possibility that members of the group are truly feigning, therebyincreasing the sensitivity. Hence, in support of some previous studies (Haines & Norris, 1995; Bianchini et al., 2001), ourresults question the research line followed since the studies by Binder at the beginning of the 1990s, identifying individ-uals in litigation as malingering suspects. The consideration of external malingering criteria in addition to the status oflitigant appears recommendable, although a consensus must be achieved on the relevant criteria to be taken into account.

4.2. Comparison with scores published in the original test manuals

There appear to be no differences in the performance of any of these tests between the Spanish and North Americansamples, with an especially high degree of similarity in the VSVT results. The means obtained for the Spanish andNorth American groups with brain damage were practically identical in all three tests, and the same was true forthe AN in the VSVT and TOMM tests (the b test manual provides no data for AN). The only statistically significantdifference observed was for the group with brain damage and high probability of malingering (G.2) in TOMM (Table 4),alongside a difference that was close to significance for the same group in the b test. Among other possible reasons, thisdifference may be explained by the heterogeneity of the groups considered to have a high probability of malingering,leading to varying percentages of malingering subjects in the different studies. The TOMM identifies patients at riskof malingering by the presence of one or more of the following factors: pursuit of a personal injury suit, pursuit ofa disability petition, and symptoms discrepant from known neurologic disease. This implies considerable differenceswith the group used in the present study.

The malingering group used by the b test comprised individuals with at least two test results indicating non-crediblecognitive symptoms, whereas the sole criteria for our group was involvement in CPS-related litigation. On the contrary,the results of their more homogeneous group did not significantly differ from those of our group, probably because of thehigh standard deviation (1,069 points) in the North American group, although the difference was close to significance(p = .070).

The differences found for the GR.2 group (Table 4) in the TOMM and b test were not produced in the VSVT, probablybecause both the designers of the test and the present authors considered involvement in litigation as sufficient to assignthe individual to the group with a high probability of malingering. In fact, when the fact of scoring as a malingerer in atleast two tests was added to the litigation criterion, producing two homogeneous groups, these differences disappeared.

According to our data, the probability of the litigants with PCS being malingerers was 50% (7/14 patients), almostidentical to the proportion of 48% reported in another study of North Americans in the same situation (Youngjonh,Burrows, & Erdal, 1995), and is very close to the 40% reported for Larrabee (2003) summing 11 studies with relevantinformation to the base rate of malingering, or to the 39% found in mild head injury reported from a survey of the Amer-ican Board of Clinical Neuropsychology membership (Mittenberg, Patton, Canyock, & Condit, 2002). Nevertheless,these results should be considered with caution given the little research on this issue. Although there are considerablesimilarities, there are also some possible differences, indicating that these tests should be applied with particular carein Spanish populations and that further cross-cultural research is warranted in forensic neuropsychology.

The sample size of in the present study was small, and these results need to be tested in a larger population,although the similarity of our findings to the North American data, including baserate of malingering, supports thegeneralizability of our results. A further limitation is that AN group was not age- and education-matched with thepatients. This was because it was decided that fourth-year psychology students would form ideal members of thisgroup, as their knowledge of psychological assessment and brain damage would allow them to elude detection morereadily. As this was not true, it would be interesting to use community-dwelling simulators, matched in age andeducation to their PSC groups, in future research.

Acknowledgements

This study has been supported with funds from the grant “Concordancia entre los aspectos clınicos, neuropsi-cologicos y de neuroimagen funcional en pacientes con sındrome post-conmocional (FIS 03/1142)”, whose principalresearcher is Dr. Manuel Gomez-Rio.


References

Bush, S. S., Ruff, R. M., Troster, J. T. B., Koffler, S. P., Pliskin, N. H., Reynolds, C. R., et al. (2005). Symptom validity assessment: Practice issuesand medical necessity. Archives of Clinical Neuropsychology, 20, 419–426.

Bender, S. D., & Rogers, R. (2004). Detection of neurocognitive feigning: Development of a multi-strategy assessment. Archives of ClinicalNeuropsychology, 19, 49–60.

Bezeau, S., & Graves, R. (2001). Statistical power effect size in clinical neuropsychology research. Journal of Clinical and Experimental Neuropsy-chology, 23, 399–406.

Bianchini, K. J., Mathias, C. W., Greve, K. W., Houston, R. J., & Crouch, J. A. (2001). Classification accuracy of the Portland Digit RecognitionTest in traumatic brain injury. The Clinical Neuropsychologist, 15, 461–470.

Binder, L. M., & Rohling, M. L. (1996). Money matters: A meta-analytic review of the effects of financial incentives on recovery after closed-headinjury. The American Journal of Psychiatry, 153, 7–12.

Boone, K. B., Lu, P., & Hezberg, D. S. (2002). The b test. Manual. Western Psychological Services.Gervais, R. O., Rohling, M. L., Green, P., & Ford, W. (2004). A comparison of WMT, CARB, and TOMM failure rates in non-head injury disability

claimants. Archives of Clinical Neuropsychology, 19, 475–487.Gouvier, W. D., Lees-Haley, P. R., & Hammer, J. H. (2003). The neuropsychological examination in the detection of malingering in the forensic

arena: Costs and benefits. In G. P. Prigatano & N. H. Pliskin (Eds.), Clinical neuropsychology and cost outcome research (pp. 405–424). NewYork: Taylor and Francis.

Greiffenstein, M. F., Baker, W. J., & Gola, T. (1994). Validation of malingered amnesia measures with a large clinical sample. PsychologicalAssessment, 6, 218–224.

Grote, C. L., Kooker, E. K., Garron, D. C., Nyenhuis, D. L., Smith, C. A., & Mattingly, M. L. (2000). Performance of compensation seeking andnon-compensation seeking samples on the Victoria Symptom Validity Test: Cross-validation and extension of a standardization study. Journalof Clinical and Experimental Neuropsychology, 22, 709–719.

Guilmette, T. J., Hart, K. J., & Giuliano, A. J. O. (1993). Malingering detection: The use of a forced-choice method in identifying organic versussimulated memory impairment. The Clinical Neuropsychologist, 7, 59–69.

Guilmette, T. J., Hart, K. J., Giuliano, A. J., & Leininger, B. E. (1994). Detecting simulated memory impairment: Comparison of the Rey Fifteen-ItemTest and the Hiscock forced-choice procedure. The Clinical Neuropsychologist, 8, 283–294.

Haines, M. E., & Norris, M. P. (1995). Detecting the malingering of cognitive deficits: An update. Neuropsychological Review, 5, 125–148.Hiscock, C. K., Branham, J. D., & Hiscock, M. (1994). Detection of feigned cognitive impairment: The two-alternative forced-choice method

compared with selected conventional tests. Journal of Psychopathology & Behavioral Assessment, 16, 95–110.Holmquist, L. A., & Wanlass, R. L. (2002). A multidimensional approach towards malingering detection. Archives of Clinical Neuropsychology,

17, 143–156.Inman, T. H., Vickery, C. D., Lamb, D. G., Edwards, C. L., & Smith, G. T. (1998). Development and initial validation of a new procedure for

evaluating adequacy of effort given during neuropsychological testing: The Letter Memory Test. Psychological Assessment, 10, 128–139.Kay, T., Harrington, D. E., Adams, R., Anderson, T., Berrol, S., Cicerone, K., et al. (1993). Definition of mild traumatic brain injury. Journal of

Head Trauma Rehabilitation, 8, 86–87.Larrabee, G. J. (2003). Detection of malingering using atypical performance patterns on standard neuropsychological tests. The Clinical Neuropsy-

chologist, 17, 410–425.Martin, R. C., Bolter, J. F., Todd, M. E., & Gouvier, W. D. (1993). Effects of sophistication and motivation on the detection of malingered memory

performance using a computerized forced-choice task. Journal of Clinical and Experimental Neuropsychology, 15, 867–880.Martin, R. C., Hayes, J. S., & Gouvier, W. D. (1996). Differential vulnerability between postconcussion self-report and objective malingering tests

in identifying simulated mild head injury. Journal of Clinical and Experimental Neuropsychology, 18, 265–275.Mittenberg, W., Patton, C., Canyock, E. M., & Condit, D. C. (2002). Base rates of malingering and symptom exaggeration. Journal of Clinical and

Experimental Neuropsychology, 24, 1094–1102.Prigatano, G. P., Smason, I., Lamb, D. G., & Bortz, J. J. (1997). Suspected malingering and the Digit Memory Test: A replication and extension.

Archives of Clinical Neuropsychology, 12, 609–619.Puente, A. E., & Perez-Garcia, M. (2000a). Psychological assessment of ethnic minorities. In G. Goldstein & M. Hersen (Eds.), Handbook of

psychological assessment (3rd ed., pp. 527–551). Amsterdam: Pergamon Press.Puente, A. E., & Perez-Garcia, M. (2000b). Neuropsychological assessment of ethnic minorities: Clinical issues. In F. Paniagua & I. Cuellar (Eds.),

Handbook of multicultural mental health (pp. 419–435). San Diego: Academic Press.Rogers, R. (Ed.). (1997). Clinical assessment of malingering and deception. New York: Guilford Press.Rogers, R., Harrell, E., & Liff, C. (1993). Feigning neuropsychological impairment: A critical review of methodological and clinical considerations.

Clinical Psychology Review, 13, 255–274.Rose, F. E., Hall, S., & Szalda-Petree, D. (1995). Portland Digit Recognition Test-computerized: Measuring response latency improves the detection

of malingering. Clinical Neuropsychologist, 9, 124–134.Slick, D., Hopp, G., Strauss, E., & Thompson, G. B. (1997). Victoria Symptom Validity Test, Version 1.0. Psychological Assessment Resources,

Inc.Slick, D. J., Hopp, G., Strauss, E., & Spellacy, F. J. (1996). Victoria Symptom Validity Test: Efficiency for detecting feigned memory impairment

and relationship to neuropsychological tests and MMPI-2 validity scales. Journal of Clinical and Experimental Neuropsychology, 18, 911–922.Slick, D. J., Hopp, G., Strauss, E., Hunter, M., & Pinch, D. (1994). Detecting dissimulation: Profiles of simulated malingerers, traumatic brain-injury

patients, and normal controls on a revised version of Hiscock and Hiscock’s Forced-Choice Memory Test. Journal of Clinical and ExperimentalNeuropsychology, 16, 472–481.


Slick, D. J., Sherman, E. M. S., & Iverson, G. L. (1999). Diagnostic criteria for malingered neurocognitive dysfunction: Proposed standards forclinical practice and research. The Clinical Neuropsychologist, 13, 545–561.

Sweet, J. J., King, F. H., Malina, A. C., Bergman, M. A., & Simmons, A. (2002). Documenting the prominence of forensic neuropsychology atnational meetings and in relevant professional journals from 1990 to 2000. The Clinical Neuropsychologist, 16, 481–494.

Tombaugh, T. N. (1996). Test of Memory Malingering (TOMM). Multi-Health System Inc..Verdejo-Garcıa, A., Alcazar, M. A., Gomez-Jarabo, G., & Perez-Garcıa, M. (2004). Pautas para el desarrollo cientıfico y profesional de la

neuropsicologıa forense. Revista de Neurologıa, 39, 60–73.World Health Organization. (1992). International Statistical Classification of Diseases and Related Health Problems (10th ed.). Geneva, Switzerland:

WHO.Youngjonh, J. R., Burrows, L., & Erdal, K. (1995). Brain damage or compensation neurosis? The controversial post-concussion syndrome. The

Clinical Neuropsychologist, 9, 112–123.

Detection of malingering in a Spanish population using three specific malingering tests

Documents

Transcript of Detection of malingering in a Spanish population using three specific malingering tests