Abstract
Greve KW, Ord JS, Bianchini KJ, Curtis KL. Prevalence of malingering in patients with chronic pain referred for psychologic evaluation in a medico-legal context.
Objective
To provide an empirical estimate of the prevalence of malingered disability in patients with chronic pain who have financial incentive to appear disabled.
Design
Retrospective review of cases.
Setting
A private neuropsychologic clinic in a southeastern metropolitan area.
Participants
Consecutive patients (N=508) referred for psychologic evaluation related to chronic pain over a 10-year period (1995–2005).
Interventions
Not applicable.
Main Outcome Measures
Prevalence of malingering was examined using 2 published clinical diagnostic systems (Malingered Pain-Related Disability and Malingered Neurocognitive Dysfunction) as well as statistical estimates based on well validated indicators of malingering.
Results
The prevalence of malingering in patients with chronic pain with financial incentive is between 20% and 50% depending on the diagnostic system used and the statistical model's underlying assumptions. Some factors associated with the medico-legal context such as the jurisdiction of a workers' compensation claim or attorney representation were associated with slightly higher malingering rates.
Conclusions
Malingering is present in a sizable minority of patients with pain seen for potentially compensable injuries. However, not all excess pain-related disability is a result of malingering. It is important not to diagnose malingering reflexively on the basis of limited or unreliable findings. A diagnosis of malingering should be explicitly based on a formal diagnostic system.
Key Words
List of Abbreviations:
CI (confidence interval), CVLT (California Verbal Learning Test), FCE (Functional Capacity Evaluation), MMPI-2 (Minnesota Multiphasic Personality Inventory-2), MND (Malingered Neurocognitive Dysfunction), MPRD (Malingered Pain-Related Disability), PDRT (Portland Digit Recognition Test), TBI (traumatic brain injury), TOMM (Test of Memory Malingering), WAIS-III (Wechsler Adult Intelligence Scale-III)MALINGERING IS A POTENTIAL problem in contexts in which there are financial incentives to appear disabled. Malingering is “the intentional production of false or grossly exaggerated physical or psychological symptoms, motivated by external incentives such as avoiding military duty, avoiding work, obtaining financial compensation, evading criminal prosecution, or obtaining drugs.”
1
(p739) Accurate estimates of the prevalence (also known as base rate and posttest odds) of pain-related malingering are necessary to inform decisions regarding individual clinical or forensic cases and to help assess the diagnostic accuracy of clinical indicators of malingering.2
Obtaining accurate estimates of prevalence may also have more general clinical and policy implications. Estimates of the prevalence of malingered disability in patients with chronic pain have varied widely, with the most reliable estimates ranging from 20% to 40%.3
, 4
, 5
, 6
In an often cited review of the detection of exaggerated pain-related disability that included base rate estimates, Fishbain et al
3
concluded that “the reviewed studies relating to the frequency of malingering within the pain setting indicate that malingering does occur and may be present in 1.25–10.4% of chronic pain patients. Because of the poor quality of the reviewed studies, it is likely that these prevalence percentages are not reliable.”(p244) In fact, the studies cited to support that estimate could not appropriately address the question of malingering prevalence, often failed to consider the incentive status of the patients, and were sometimes irrelevant to chronic pain. Moreover, some were simply misrepresented. For example, the studies by Miller,, which were represented as examinations of patients with chronic pain, actually studied patients with head injuries. Fishbain3
also stated that 48 patients out of a cohort of 4000 were identified as malingering. This is the 1.25% that represented the lower bound of their prevalence estimate. In fact, Miller, examined 200 consecutive cases drawn from that larger medico-legal cohort and determined that 47 (not 48) of those were “unequivocally psychoneurotic complaints.”(p919) Use of the correct numerator (47 not 48) and denominator (200 not 4000) results in a malingering prevalence of 23.5%. Thus, Fishbain's3
range of estimated malingering prevalence should have been from about 10% to 24%, rather than 1% to 10%. In short, given the multitude of problems with the data, the Fishbain3
findings should not be considered a scientifically valid estimate of the prevalence of malingering in chronic pain.There are other data, however, that speak to the potential prevalence of malingering generally and in chronic pain. A nationwide survey found that 20% of Americans, and 46% of those living in 3 northeast states, felt that purposeful misrepresentation of claims in the compensation system was acceptable.
9
, 10
Consistent with this finding, 20% of patients with compensation claims undergoing treatment in local pain clinics showed indications of malingering on covert video surveillance.4
A survey of board certified neuropsychologists estimated the frequency of malingering in patients with chronic pain with financial incentive (ie, those involved in personal injury litigation or workers' compensation) at slightly over 30%.6
Thus, these studies suggest that the prevalence of malingering in chronic pain ranges from 20% to 40%. These estimates of malingering in pain are consistent with estimates in other patient populations (eg, traumatic brain injury, toxic exposure) that were based on direct assessment of patients (rather than a survey of practitioners).11
, 12
However, no studies have reported estimates of the prevalence of malingering in chronic pain based on direct, individual examination using a consistent assessment approach and formal diagnostic system.Purpose
The explicit purpose of this study was to estimate directly the base rate of malingering in patients with chronic pain through the careful examination of archival case files. Data were obtained from the files of 508 persons referred for psychologic evaluation related to chronic pain over a 10-year period. All had financial incentive, usually in the form of a workers' compensation claim or personal injury lawsuit. This study used 2 methodologically distinct but somewhat overlapping approaches to estimate malingering rates: clinical diagnosis using published diagnostic systems and statistical estimation based on psychometric test performance. Both methods were used to estimate the prevalence of malingering across the entire sample and within specific patient subgroups. The statistical method also allowed a comparison of rates of malingering based on the different indicator types.
The statistical method differs from the clinical method in that it does not identify individual malingerers but estimates the proportion of the sample who would likely have been classified as malingering based on the clinical systems. However, some overlap between the 2 methods occurs because the malingering indicators used in the statistical approach also contribute partially, but not necessarily exclusively, to a diagnosis of malingering within the clinical classification systems. The use of 2 complementary methods potentially provides some additional validation of the estimates. This study thus offers unique information about the rate of malingering in chronic pain and about factors that may influence that malingering rate.
Methods
Participants
This study was conducted according to the guidelines for the protection of human participants. Data were abstracted from the files of 508 consecutive cases referred for psychologic evaluation related to chronic pain in the context of identifiable financial incentive. All data were archival and collected from 1995 to 2005 in the course of the clinical psychology practice in a single southeastern metropolitan area. Patients were included if they reported pain-related disability regardless of injury type, location, or etiology. Most patients were seen for a primary complaint of back pain. Most of the patients with a primary back/spine injury had no objective evidence of pathology involving the spine or spinal cord. Although not tabulated, almost all the patients were being treated with a variety of medications, many had undergone injections, and most had received physical therapy.
Fifty-five percent of the cases were represented by an attorney; 20% were not represented; and 25% did not indicate representation status. At the time of the evaluation, approximately 89% were involved in workers' compensation claims, 9% in personal injury litigation, and 2% in disability claims. Note that it is likely that some workers' compensation claimants were also involved in related litigation outside the workers' compensation system. Of the workers' compensation claimants, 70% were covered under state workers' compensation law, while 19% were covered under federal regulations. Table 1 presents the descriptive statistics for the sample as a whole and the nature of the financial incentive. Table 2 summarizes the pathology and surgical history of the groups. Table 3 presents features of the medico-legal context.
In patients with primary spine injury.
Table 1Sample Demographic Characteristics (N=508)
| Variable | n | Mean ± SD |
|---|---|---|
| Age (y) | 508 | 42.1±9.6 |
| Education (y) | 508 | 11.6±2.6 |
| Time since injury (mo) | 506 | 37.1±38.7 |
| n | % | |
| Sex | ||
| Men | 328 | 64.6 |
| Women | 180 | 35.4 |
| Race | ||
| White | 325 | 64.0 |
| Black | 135 | 26.6 |
| Hispanic | 10 | 2.0 |
| Native American | 1 | 0.2 |
| Other | 2 | 0.4 |
| Not indicated | 35 | 6.9 |
Table 2Injury/Symptom Characteristics of the Chronic Pain Sample
| Characteristics | % |
|---|---|
| Injury location | |
| Primary back/spine injury | 79.5 |
| Head injury in accident | 12.4 |
| Symptom location | |
| Head | 26.3 |
| Spine | 90.3 |
| Chest/abdomen | 8.1 |
| Upper extremity | 41.6 |
| Lower extremity | 66.7 |
| Spine findings | |
| Any spine findings | 20.6 |
| Degenerative disk/spine | 7.8 |
| Herniated nucleus pulposus | 4.7 |
| Disk bulge/protrusion | 8.6 |
| Neural impingement | 1.8 |
| Spinal surgery | |
| Diskectomy/fusion | 35.4 |
| Decompression/laminectomy | 16.4 |
| Other pain diagnoses | |
| Complex regional pain syndrome | 4.8 |
| Fibromyalgia | 3.5 |
| Myofascial pain syndrome | 3.5 |
NOTE. Percentage value represents the percentage of patients within a group with the indicated presentation, symptom, finding, or procedure. Note that an individual patient may be positive in more than 1 category, so the sum of percentages in a particular category may be greater than 100.
Table 3Medico-Legal Characteristics of the Sample
| Characteristic | n | % |
|---|---|---|
| Referral source | ||
| Doctor | 169 | 33.3 |
| Case manager/adjuster | 249 | 49.0 |
| Attorney | 90 | 17.7 |
| Claim type | ||
| Disability | 8 | 1.6 |
| Workers' compensation | 452 | 89.0 |
| Personal injury | 48 | 9.4 |
| Represented by an attorney | ||
| No | 103 | 20.3 |
| Yes | 277 | 54.5 |
| Not indicated | 128 | 25.2 |
| Workers' compensation source | ||
| State | 357 | 70.3 |
| Federal | 95 | 18.7 |
Measures
Indicators of malingering used in this study can be divided into 3 types related to the assessment method:
- 1Stand-alone forced-choice symptom validity tests that appear to assess memory but that really assess whether adequate effort to perform well was provided. These include the PDRT,13,14the TOMM, and the Word Memory Test.16
- 2Internal or embedded indicators derived from standard tests of cognitive ability. These include the Reliable Digit Span,17the Digit Span scaled score, Processing Speed and Working Memory Indexes from the WAIS-III18and Recognition Hits, the Millis Formula,19and the Linear Shrinkage Model20from the CVLT.21,22
- 3Measures of symptom exaggeration and overreporting derived from the MMPI-2.23
Procedure
Clinical classification methods
The clinical malingering classification method applies systemized criteria for the diagnosis of malingering to each patient. Estimates were made using 2 diagnostic systems: the criteria by Bianchini et al
24
for MPRD, and the criteria by Slick et al25
for MND. Both sets of criteria have been used extensively in research and clinical practice and have served as the method for operationalizing malingering in numerous published studies of malingering in a variety of compensable psychologic, neuropsychologic, and medical conditions, including chronic pain.26
Appendices 1 and 2, Appendices 1 and 2 summarize the MND and MPRD criteria, respectively.Each patient's cognitive malingering (MND) status was first classified using a diagnostic decision tree.
27
In determining the presence of MND, each case was evaluated on the basis of 4 criteria: (A) presence of substantial external incentive, (B) evidence from neuropsychologic testing, (C) evidence from self-report, and (D) behaviors meeting the necessary B and C criteria not fully accounted for by psychiatric, neurologic, or developmental factors. Using this system, all diagnoses of malingering required the presence of an external incentive (criterion A) plus criterion B and/or C evidence (see appendix 1) in the absence of explanatory factors (criterion D).MND classification relied entirely on psychometric indicators of malingering. Cutoffs for these indicators (table 4) were based on examination of classification accuracy data derived from published known-groups TBI studies and in consideration of the general literature on specific indicators. The published source
28
, 29
, 30
, 31
, 32
, 33
of each cutoff is noted in table 4 along with the associated classification accuracy. Two additional forced-choice measures, the Word Memory Test and the Computerized Assessment of Response Bias, were also used to identify below-chance responding (criterion B1 and C1 for the MND and MPRD systems, respectively).Table 4Classification Criteria for Psychometric Variables Used in the Clinical Diagnosis of MND and MPRD
| Positive | Ambiguous | Below Chance | Criterion | ||||
|---|---|---|---|---|---|---|---|
| Indicator | Cutoff | FP (%) | Cutoff | FP (%) | Cutoff | MND | MPRD |
| PDRT 28 | |||||||
| Easy | <23 | 1 | <27 | 17 | <12 | B1/B2 | C1/C2 |
| Hard | <18 | 2 | <23 | 14 | <12 | B1/B2 | C1/C2 |
| Total | <45 | 2 | <50 | 19 | <28 | B1/B2 | C1/C2 |
| TOMM 29 | |||||||
| Trial 2 | <45 | 4 | <49 | 14 | <18 | B1/B2 | C1/C2 |
| Retention | <45 | 3 | <49 | 11 | <18 | B1/B2 | C1/C2 |
| WAIS-III 30 , 31 , 32 | |||||||
| RDS | <7 | 4 | <8 | 28 | NA | B2 | C2 |
| DS | <5 | 0 | <7 | 24 | NA | B6 | C3 |
| WMI | <76 | 14 | <81 | 27 | NA | B6 | C3 |
| PSI | <71 | 11 | <76 | 31 | NA | B6 | C3 |
| MMPI-2 33 | |||||||
| F | >80 | 4 | >65 | 27 | NA | C5 | D5 |
| Fb | >85 | 3 | >65 | 28 | NA | C5 | D5 |
| Fp | >80 | 2 | >65 | 12 | NA | C5 | D5 |
| FBS | >27 | 6 | >24 | 23 | NA | C5 | D5 |
| Meyers | >6 | 2 | >3 | 20 | NA | C5 | D5 |
NOTE. These FP values were derived from independent traumatic brain injury samples not for the patients with chronic pain examined in this study.
Abbreviations: DS, Digit Span; F, infrequency; F(b), infrequency (back); FBS, Symptom Validity Scale; F(p), infrequency (psychopathology); FP, false-positive error rate; Meyers, Meyers composite validity index for the MMPI-2; NA, not applicable; PSI, Processing Speed Index; RDS, Reliable Digit Span; WMI, Working Memory Index.
The MPRD
24
system, developed specifically for use in chronic pain populations, evaluated each case on 5 criteria: (A) evidence of significant external incentive, (B) evidence from physical evaluation, (C) evidence from cognitive/perceptual (neuropsychologic) testing, (D) evidence from self-report, and (E) behaviors meeting necessary criteria from groups B, C, and D not fully accounted for by psychiatric, neurologic, or developmental factors. In this system, a diagnosis of malingering required the presence of external incentive (criterion A) plus evidence from criterion B, C, and/or D (see appendix 2) in the absence of explanatory factors (criterion E).Two methods were used to operationalize the MPRD criteria, with the primary difference being in the criteria used to establish probable malingering. The first method (referred to here as MPRD−) uses only psychometric indicators to diagnosis probable malingering. MPRD− differs from the MND criteria only in terms of how test findings can be combined to reach a diagnosis of probable malingering (ie, the MND criteria require at least 1 finding consistent with cognitive malingering). Although both the MND and MPRD systems allow qualitative findings (inconsistencies) to contribute to a diagnosis, in this study, neither MND nor MPRD− used such inconsistencies and instead relied solely on psychometric findings. In contrast, the second approach to diagnosing MPRD (MPRD+) used psychometric indicators in addition to a variety of qualitative inconsistencies. Psychometric cutoffs for both MPRD systems were applied as noted in table 4.
Four kinds of inconsistencies were considered as part of the MPRD+ classification: (1) nonorganic or functional findings on physical examination (exclusive of FCE), (2) an inconsistency between the patients' behavior during examination and their behavior when they did not believe they were being observed, (3) inconsistencies between the patients' subjective report of symptoms or history and their documented history, and (4) evidence of submaximal effort, symptom magnification, or nonorganic/functional findings on a formal FCE. To account for their qualitative nature, this study required that at least 2 documented inconsistencies be present to meet criteria and contribute to a diagnosis of MPRD+. The 1 exception was in the case of a “compelling inconsistency.”
24
(p408)A compelling inconsistency occurs when the difference in the way patients present when being evaluated compared with when they are not aware of being evaluated is so inconsistent that it is not reasonable to believe patients are not purposely controlling the difference.
24
Patients coded as having compelling inconsistencies were examined in detail, and 2 of the authors (K.W.G., K.J.B.) had to agree that the inconsistency was compelling as defined. If there was no agreement, then the inconsistency was considered a normal qualitative inconsistency and treated as described.Table 4 shows how each psychometric indicator used in this study was specifically used within the MND and MPRD systems. In order to describe the sample better, patients were grouped into 1 of 6 categories using each classification system: (1) no positive findings, (2) 1 ambiguous finding, (3) multiple ambiguous findings, (4) 1 positive finding, (5) probable malingering, or (6) definite malingering. Consistent with the overall malingering literature, patients meeting criteria for definite and probable are considered to be malingering and represent the basis of the prevalence calculation for each diagnostic system.
26
, 27
The malingering detection literature combines these 2 groups for 2 reasons: (1) they are essentially indistinguishable in terms of overall malingering findings, and (2) from a medico-legal standpoint (which is particularly relevant here), both meet the standard of “more probable than not” or “to a reasonable degree of scientific certainty.” Note that ambiguous findings did not contribute to a diagnosis of malingering in either system.Statistical estimation method
The second method identifies malingering in a statistically probabilistic manner using parameter estimates (sensitivity and specificity) derived from malingering detection research in TBI. This approach is conservative because TBI samples include patients with objectively demonstrable brain damage. Indicators from the PDRT,
28
TOMM,29
WAIS-III,30
, 31
, 32
CVLT,34
and MMPI-233 were used, because published classification accuracy data were available for a range of cutoffs derived from the performance of patients with TBI.Estimating the base rate of malingering in this way requires the use of positive and negative predictive power. Predictive power is an index of the probability that test result is accurate. That is, positive predictive power (measured: true-positives divided by the sum of true-positives and false-positives) is the probability that a positive test result was produced by a person with the given condition (eg, malingering). In contrast, negative predictive power (measured: true-negatives divided by the sum of true-negatives and false-positives) is the probability that a negative finding was produced by a person without the condition.
Predictive power integrates both sensitivity and specificity. Moreover, when applied to the group of patients who had a positive result on a given test, positive predictive power provides an index of the proportion of those persons who would likely meet the criteria for malingering on which the original sensitivity and specificity data were based. Similarly, when negative predictive power is applied to the group of patients who were negative, it reflects the proportion of that group who would not have met criteria for malingering. Subtracting this value from 1 leaves the proportion of persons who had a negative result on the test but who would have met criteria for malingering. In short, predictive power corrects for misclassifications (ie, false-positives and false-negatives) using the known error rates of an indicator.
When the values derived with positive predictive power are combined with those derived from negative predictive power, one has an estimate of the number or proportion of true malingerers. Appendix 3 summarizes this process.
The complication with this seemingly simple statistical approach is that the calculation of predictive power requires an explicit assumption about pretest odds.
36
For the purposes of this study, we have modeled the base rate of malingering in pain using 3 different pretest odds (.20, .30, .40). These estimates are supported by (1) observed base rates of malingering in similar populations with financial incentive,12
(2) survey6
prevalence estimates, and (3) results from the clinical classification method used in the present study, which are made independently of pretest odds.Predictive power was estimated for all scores based on data cited in the studies noted for cutoffs associated with a specificity of approximately 95% in TBI populations. Because of the features of some score distributions, cutoffs were not always available at exactly the 95% specificity level. Cutoffs and precise specificity and sensitivity values used for predictive power calculations for each indicator are presented with the analysis results. Note that the indicators and cutoffs used for statistical modeling are slightly different from those used in the clinical diagnostic process because of the ability of the statistical method to incorporate both sensitivity and specificity parameters.
Results
Clinical Classification Methods
Of the 508 patients, probable or definite malingering was observed in 25.2% (95% CI, 21.4–29.0; n=128) cases using the criteria by Slick et al
25
for MND, 32.5% (95% CI, 28.4–36.6; n=168) cases using MPRD− criteria, and 36% (95% CI, 31.8–40.2; n=183) cases using MPRD+ criteria. Below-chance responding on a forced-choice symptom validity test was observed in 8.9% of the sample, and compelling inconsistencies in patient report were observed in an additional 1.5%, resulting in 10.4% (95% CI, 7.8–13.1) of the sample being classified as definite malingering. About one third of the sample had some findings consistent with symptom exaggeration/malingering but did not meet criteria for a diagnosis of probable malingering. In the system by Bianchini et al,24
these may be classified as possible MPRD. Table 5 presents the proportions of patients falling into each malingering classification for each system. MPRD+ status was not meaningfully correlated with age (r=.017; r2<.001) or time since injury (r=.043; r2=.002), and no significant differences in prevalence rates were observed between males and females (χ2=.875; df=1; P=.349). MPRD+ status was significantly correlated with education, although the relationship was relatively weak (r=–.250; r2
=.063).Table 5Proportions of the Sample Meeting Levels of Malingering Criteria According to Each Classification System
| Malingering Status | MND | MPRD− | MPRD+ |
|---|---|---|---|
| n(%) | n(%) | n(%) | |
| Not malingering | |||
| No indicators of malingering | 76(15.0) | 61(12.0) | 58(11.4) |
| One ambiguous finding | 71(14.0) | 37(7.3) | 34(6.7) |
| Multiple ambiguous findings | 71(14.0) | 93(18.3) | 89(17.5) |
| One positive finding | 162(31.9) | 152(29.9) | 144(28.3) |
| Malingering | |||
| Probable malingering | 83(16.3) | 112(22.0) | 130(25.6) |
| Definite malingering | 45(8.9) | 53(10.4) | 53(10.4) |
| Totals (N=508) | |||
| Not malingering | 380(74.8) | 343(67.5) | 325(64.0) |
| Malingering | 128(25.2) | 165(32.5) | 183(36.0) |
Statistical Estimation Method
It should be noted that because of some changes in the composition of the test battery over time, not all patients received exactly the same tests. The number and percentage of patients who had scores on each test are reported in table 6 along with the percentage of positive findings on each test score examined and the related prevalence estimate. Note that the prevalence is modeled using predictive power, which assumes a certain posttest odds. Prevalence was modeled with 3 different posttest odds assumptions. The data in table 6 represent the base rate of malingering using single indicators alone, and these values reflect the performance of 450 to 508 cases, depending on the specific test or indicator. Percentages of positive findings on each indicator range from 10.5% (Digit Span) to 34% (Millis formula from the CVLT). Prevalence estimates based on these values range from 19.1% to 51.3% depending on the pretest odds used. Overall, prevalence estimates based on cognitive indicators were approximately 10% less than estimates based on indicators of exaggerated somatic symptoms.
Sensitivity and false-positive error rates are based on known-groups calibration studies in TBI samples.
Table 6Cutoffs, Calibrated Classification Accuracy, Rate of Positive Test Results, and Estimated Malingering Prevalence for Each Indicator
| Indicator | TBI Calibration | Current Sample | Estimated Prevalence | |||||
|---|---|---|---|---|---|---|---|---|
| Cutoff | FP | Sensitivity | n | Pos % | .20 | .30 | .40 | |
| Forced-choice symptom validity tests | ||||||||
| PDRT 28 | ||||||||
| Easy | <25 | 5 | 66 | 501 | 21.8 | 23.2 | 26.6 | 34.7 |
| Hard | <19 | 5 | 52 | 501 | 18.0 | 22.2 | 27.4 | 36.5 |
| Total | <47 | 5 | 70 | 501 | 26.3 | 26.0 | 28.5 | 36.5 |
| TOMM 29 | ||||||||
| Trial 1 | <39 | 5 | 54 | 451 | 17.7 | 21.7 | 27.1 | 35.7 |
| Trial 2 | <47 | 5 | 49 | 451 | 14.6 | 20.6 | 26.4 | 35.0 |
| Retention | <47 | 5 | 61 | 450 | 14.9 | 19.1 | 23.6 | 31.6 |
| Internal validity indicators | ||||||||
| CVLT 34 | ||||||||
| Recognition hits | <11 | 6 | 47 | 476 | 13.9 | 20.0 | 26.1 | 35.1 |
| Millis formula | <–1 | 3 | 26 | 476 | 34.0 | 33.8 | 39.7 | 51.3 |
| Linear shrinkage | >1 | 6 | 56 | 476 | 21.6 | 23.3 | 28.2 | 37.4 |
| WAIS-III 30 , 31 , 32 | ||||||||
| Reliable Digit Span | <7 | 4 | 39 | 508 | 12.6 | 20.9 | 27.6 | 36.8 |
| Digit Span | <6 | 7 | 36 | 486 | 10.5 | 19.1 | 26.8 | 36.2 |
| Psychologic test validity indicators | ||||||||
| MMPI-2 33 | ||||||||
| F | >80 | 4 | 30 | 508 | 20.1 | 25.4 | 32.5 | 42.9 |
| F(b) | >80 | 4 | 47 | 508 | 29.7 | 30.7 | 35.0 | 45.3 |
| F(p) | >75 | 5 | 21 | 508 | 11.8 | 21.3 | 29.9 | 40.2 |
| FBS | >28 | 3 | 41 | 508 | 28.5 | 31.5 | 36.0 | 46.5 |
| Meyers Index 61 | >5 | 6 | 48 | 507 | 27.6 | 27.2 | 32.5 | 42.8 |
| Hysteria | >89 | 7 | 29 | 508 | 31.5 | 27.0 | 34.7 | 46.1 |
| Hypochondriasis | >89 | 6 | 25 | 508 | 24.8 | 25.2 | 33.1 | 44.5 |
Abbreviations: F, infrequency; F(b), infrequency (back); FBS, Symptom Validity Scale; F(p), infrequency (psychopathology); FP, false-positive error rate; Pos, positive for malingering.
† These values are the hypothetical base rates on which statistical malingering prevalence was modeled.
Table 7 presents breakdowns of positive findings and prevalence estimates averaged across all indicators, by indicator type method, and by behavioral domain. Note that these values are averages of estimates for each indicator which are themselves based on about 500 cases. As seen in table 7, on average approximately 21% of the sample was positive across all the examined indicators. When the classification accuracy of each indicator and population estimates are taken into account, the estimated prevalence of malingering ranged from 24% (95% CI, 20.3–27.7) to 40% (95% CI, 35.7–44.3), rates in close agreement with the clinical diagnostic methods.
These values are the hypothetical base rates on which statistical malingering prevalence was modeled.
Table 7Mean Rate of Positive Test Results and Estimated Malingering Prevalence for All Indicators and by Indicator Type and Domain
| Indicator Breakdown | Pos (%) | .20 | .30 | .40 |
|---|---|---|---|---|
| All indicators | ||||
| All indicators of response bias (n=18 variables) | 21.1±7.3 | 24.3±4.4 | 30.1±4.3 | 39.7±5.4 |
| Indicator type | ||||
| Forced-choice symptom validity tests (n=6 variables from the PDRT, TOMM) | 18.9±4.5 | 22.1±2.3 | 26.6±1.7 | 35.0±1.8 |
| Internal validity indicators (n=5 variables from the CVLT, WAIS) | 18.5±9.6 | 23.4±6.0 | 29.6±5.7 | 39.4±6.7 |
| Psychologic test validity indicators (n=7 variables, F, F(b), F(p), FBS, HS, HY) | 24.9±6.8 | 26.9±3.5 | 33.4±2.0 | 44.0±2.2 |
| Domain | ||||
| Cognitive validity indicators (n=11 variables, PDRT, TOMM, CVLT, WAIS) | 19.7±6.9 | 22.7±4.2 | 28.0±4.1 | 37.0±5.0 |
| Psychiatric symptom validity indicators (n=3 variables, F, F(b), F(p)) | 20.5±9.0 | 25.8±4.7 | 32.5±2.6 | 42.8±2.6 |
| Somatic symptom validity indicators (n=3 variables FBS, HS, HY) | 28.3±3.4 | 27.9±3.2 | 34.6±1.5 | 45.7±1.0 |
NOTE. Values are mean ± SD.
Abbreviations: F, infrequency; F(b), infrequency (back); FBS, Symptom Validity Scale; F(p), infrequency (psychopathology); FP, false-positive error rate; Hs, hypochondriasis; Hy, hysteria; n, the number of variables or scores contributing to the mean; each mean is a composite of individual variable means which themselves are a composite of about 500 cases (see table 6); Pos, average percentage of the entire sample scoring in the positive range for each indicator or combination of indicators; Sens, sensitivity.
Table 7 also presents hit rate and prevalence rate data averaged across the 3 different types of indicators: (1) stand alone forced-choice symptom validity tests that appear to assess memory but which really assess effort expended to do well (ie, PDRT, TOMM), (2) indicators that are derived from standard tests of cognitive ability (in this case, the WAIS and CVLT), and (3) measures of symptom exaggeration and overreporting derived from self-report questionnaires (ie, MMPI-2). These data suggest that the different indicator types produced comparable estimates of prevalence that did not appear to differ from the summary values, although estimates based on indicators of symptom exaggeration were slightly higher than those based on indicators of cognitive underperformance.
Finally, table 7 also presents hit rate and prevalence rate data averaged across 3 behavioral domains in which underperformance or exaggeration may occur: (1) cognitive underperformance, (2) exaggeration of psychiatric symptoms, and (3) exaggeration of somatic symptoms. Again, the estimates for each behavioral domain closely track the overall estimates, and one can see the tendency toward a higher rate of exaggeration of somatic symptoms, especially compared with cognitive symptoms. This is not surprising given that the patients in this study were being evaluated primarily in relation to back injuries and present primarily with physical (as opposed to cognitive) complaints.
Contextual factors
The association of malingering with different features of the medico-legal context was also examined. Specific factors examined included referral source (doctor, case manager/adjuster, attorney), claim type (workers' compensation, personal injury litigation, disability claim), attorney representation, and type of workers' compensation case (state, federal). Overall, the statistically based estimates within these subgroups were comparable to the overall estimates discussed, and they were similar to the rates of clinically diagnosed MND. This is not surprising, because both the statistical approach and MND diagnosis (as operationalized here) relied entirely on psychometric indicators. The rates of MPRD diagnoses demonstrated some differences among subgroups, with the largest differences seen for the MPRD+ diagnosis. MPRD+ rates were highest among those represented by an attorney (39.4%) and those involved in workers' compensation claims (36.7%), especially federal workers' compensation (40%). However, chi-square analysis indicated that none of the differences in MPRD+ malingering rates according to these contextual breakdowns were significant (P>.05). Table 8 presents a full breakdown for each clinical classification system and rates based on statistical estimation.
Table 8Proportion of Patients Meeting Criteria for Malingering Using Clinical Classification and Statistical Estimation as a Function of Medico-Legal Contextual Factors
| Clinical Diagnosis | Statistical Estimation | ||||||
|---|---|---|---|---|---|---|---|
| Contextual Factor | MND | MPRD− | MPRD+ | Pos (%) | .20 | .30 | .40 |
| Referral source | |||||||
| Doctor | 20.7 | 29.6 | 30.9 | 18.6 | 22.8 | 28.8 | 38.2 |
| Case manager | 27.7 | 32.9 | 37.3 | 22.7 | 25.2 | 31.0 | 40.6 |
| Attorney | 26.7 | 36.7 | 38.9 | 21.7 | 24.8 | 30.4 | 39.9 |
| Claim type | |||||||
| Disability | 14.3 | 28.6 | 28.6 | 5.6 | 19.2 | 22.4 | 31.6 |
| Workers' compensation | 25.7 | 32.7 | 36.7 | 21.7 | 24.6 | 30.3 | 40.1 |
| Personal injury | 22.9 | 31.3 | 31.3 | 18.6 | 22.8 | 29.0 | 38.0 |
| Represented by an attorney | |||||||
| No | 21.4 | 28.2 | 31.1 | 18.9 | 23.1 | 28.9 | 38.5 |
| Yes | 28.2 | 36.1 | 39.4 | 22.6 | 25.1 | 30.8 | 40.6 |
| Not indicated | 21.9 | 28.1 | 32.8 | 19.7 | 23.5 | 29.4 | 38.8 |
| Workers' compensation source | |||||||
| State | 25.8 | 31.9 | 35.9 | 21.7 | 24.6 | 30.3 | 40.0 |
| Federal | 25.3 | 35.8 | 40.0 | 21.6 | 24.6 | 30.2 | 40.0 |
Abbreviation: Pos, average percent positive across all indicators.
Discussion
The present study presents estimates of the prevalence of malingering in patients with chronic pain referred for psychologic evaluation. Previously published estimates of malingering rates in chronic pain have suffered from a variety of methodologic and conceptual limitations. This study is the first to use explicit operationalization of malingering and data from direct assessment of patients to estimate malingering prevalence in patients with chronic pain. The results of this study suggest that the prevalence of malingering in patients with pain with financial incentive is between 20% and 50%, depending on specific methods and assessment context. The rates of malingering observed in this study are consistent with methodologically sound survey results
6
and are generally in line with the findings from other clinical conditions.11
, 12
, 37
, 38
, 39
Having a workers' compensation claim, especially in a federal jurisdiction, and being represented by an attorney were associated with slightly higher rates of malingering. Overall, these results emphasize the fact that malingering is present in a sizable minority of patients with pain seen for potentially compensable injuries.Study Limitations
Certain methodologic factors also have an effect on the magnitude and precision of the estimates reported here. For example, the malingering diagnostic system used influenced prevalence estimates. This is not surprising. Estimates based on MND criteria
12
would be expected to underestimate malingering in patients with pain because this system is weighted toward evidence of exaggerated cognitive deficits. In the absence of evidence of cognitive exaggeration, MND cannot be diagnosed. In contrast, the MPRD24
criteria were developed for use in chronic pain and gave equal weight to signs of cognitive, psychologic, and physical malingering. The MPRD system classified more patients as malingering than the MND system. The use of both of these clinical diagnostic systems resulted in base rate estimates ranging from 25% to 35%. The statistical method resulted in slightly higher estimates.Another factor influencing the malingering prevalence estimates reported in this study was the decision to use cutoffs for psychometric indicators that were derived from TBI samples. In both the clinical classification and statistical estimation models, the application of these cutoffs should be considered conservative. That is, the criteria for a malingering diagnosis and the low false-positive error rates associated with the cutoffs selected for individual indicators means that the probability that nonmalingerers have been incorrectly classified as malingering is very low. The TBI samples used to calibrate the various indicators included patients with objectively demonstrable brain pathology and dysfunction. Thus, when applying these cutoffs, which are conservative even in brain-injured populations, to chronic pain, the rate of false-positive errors would be expected to be even lower than originally seen in TBI. This is particularly true for indicators based on cognitive performance. Related to this, the sensitivity of these measures may also be lower because patients with chronic pain may have less reason specifically to exaggerate cognitive (and to some extent psychiatric) symptoms. Taken together, the chosen indicators and the selected cutoffs were much more likely to err on the side of failing to detect malingerers. Thus, the use of these cutoffs in this study was more likely to result in an underestimation of the rate of malingering in chronic pain.
In general, the accuracy of estimates of prevalence are heavily dependent on sampling methodology and sample size; larger samples are more likely to be representative of the population of interest than smaller samples.
40
The present sample was large, but some of the subsamples were relatively small (eg, the disability and personal injury cases), so the margin of error will tend to be larger in those subgroups. As to representativeness, the present sample reflects all chronic pain cases with incentive seen in an active clinical psychology group practice over a 10-year period. It is arguable that this sample is representative of the kinds of chronic pain patients typically seen for psychologic evaluation in the Southeast Louisiana region. The fact that the present findings are in line with those of a national survey6
suggest that the present estimates can be generalized. However, because the current study included only patients with pain referred for psychologic evaluations, the prevalence of malingering may be greater than in patients seen in general medical, specialty medical, or rehabilitation contexts.Neuropsychology has spearheaded research on the detection of malingering,
24
which resulted in a heavy emphasis on the detection of cognitive malingering (see recent edited texts26
, 41
). Consequently, there is a relative lack of well validated tools for detecting intentional exaggeration and underperformance outside the cognitive domain. Some indicators of exaggerated subjective psychologic and somatic symptoms have been developed and validated, but not with the variety seen in cognitive indicators. Measures of physical capacity (eg, finger tapping speed, grip strength, other components of FCE) will likely be of particular value in chronic pain, and some development of these tools has been reported.42
Of particular promise are the kinds of procedures often included in FCEs. These types of procedures are sensitive to volitional effort and may thus be capable of identifying nonphysiologic performance patterns including malingering.43
, 44
, 45
, 46
, 47
, 48
, 49
The precision of malingering detection in chronic pain will improve as methods for directly assessing the validity of symptoms and performance are refined.Conclusions
This article has demonstrated that the prevalence of malingering is higher than may be commonly appreciated. However, while it is clear that most patients with chronic pain do not meet criteria for malingering, nearly half the sample showed some evidence of symptom magnification, and a third met criteria for possible MPRD. As many as two thirds of our patients show some form of exaggeration. About half of the nonmalingering patients with back injury complaining of pain in 3 or more areas of their body had objective physical findings (excluding surgery). In contrast, 21% of malingerers had objective physical findings. That is, patients with pain classified as malingering had more pain complaints with less objective underlying pathology than did the nonmalingerers.
When there is less objective evidence of pathology present in the context of significant pain-related disability, one should be more concerned about psychologic involvement including malingering. The finding of excess disability or other inconsistencies should trigger a more detailed analysis of psychologic causality. Psychosocial factors, including somatization, are a high prevalence phenomenon in primary care
50
and neurology practices51
and have important implications for outcome in chronic pain.52
, 53
, 54
, 55
The formal examination of psychologic causality (via both formal psychologic evaluations and quantitative physical capacity examinations) is important because under most circumstances, psychologic overlay and malingering cannot be reliably identified in the clinical evaluation in the absence of specialized tests.
56
, 57
, , 59
, 60
The fact that much of the excess disability seen in patients with pain reflects complicating psychologic factors emphasizes the need for care in the use of the term malingering. Not all exaggeration reflects malingering. Malingering should not be diagnosed reflexively on the basis of limited or unvalidated but suspicious findings, and we urge the use of explicit criteria for the diagnosis of malingering like those of Bianchini et al24
and Slick et al.25
Appendix 1: Summary of the Slick, Sherman, & Iverson25 Criteria for MND
- APresence of substantial external incentive
- BEvidence from neuropsychologic testing
- 1Definite negative response bias
- 2Probable response bias
- 3Discrepancy between test data and known patterns of brain functioning
- 4Discrepancy between test data and observed behavior
- 5Discrepancy between test data and reliable collateral reports
- 6Discrepancy between test data and documented background history
- 1
- CEvidence from self-report
- 1Self-reported history is discrepant with documented history
- 2Self-reported symptoms are discrepant with known patterns of brain functioning
- 3Self-reported symptoms are discrepant with behavioral observations
- 4Self-reported symptoms are discrepant with information obtained from collateral informants
- 5Evidence from exaggerated or fabricated psychologic dysfunction
- 1
- DBehaviors meeting necessary criteria from groups B and C are not fully accounted for by psychiatric, neurologic, or developmental factors
Diagnostic Categories for Malingered Neurocognitive Dysfunction
- IDefinite
- 1Presence of substantial external incentive (criterion A)
- 2Definite negative response bias (criterion B1)
- 3Behaviors meeting necessary criteria from groups B and C are not fully accounted for by psychiatric, neurologic, or developmental factors (criterion D)
- 1
- IIProbable
- 1Presence of substantial external incentive (criterion A)
- 2Two or more types of probable evidence of intent from B criteria (B2–B6) or one B criterion (B2–B6) and one or more C criteria
- 3Behaviors meeting necessary criteria from groups B and C are not fully accounted for by psychiatric, neurologic, or developmental factors (criterion D)
- 1
- IIIPossible
- 1Presence of substantial external incentive (criterion A)
- 2Evidence does not rise to the level sufficient for a diagnosis of Probable MPRD➢Meets only one B criterion (B2–B6)OR➢Meets one or more C criteriaOR➢Evidence sufficient for a diagnosis of MPRD is present but criterion E is not met
- 1
Appendix 2: Summary of the Criteria for the Diagnosis of MPRD24
- AEvidence of significant external incentive
- BEvidence from physical evaluation
- 1Probable effort bias
- 2Discrepancy between subjective report of pain and physiologic reactivity
- 3Nonorganic findings
- 4Discrepancy between the patient's physical presentation during formal evaluation and physical capacities documented when they are not aware of being observed
- 1
- CEvidence from cognitive/perceptual (neuropsychologic) testing
- 1Definite negative response bias
- 2Probable response bias
- 3Discrepancy between cognitive/neuropsychologic test data and known patterns of brain functioning
- 4Discrepancy between test data and observed behavior
- 1
- DEvidence from self-report
- 1Compelling inconsistency
- 2Self-reported history is discrepant with documented history
- 3Self-reported symptoms are discrepant with known patterns of physiologic or neurologic functioning
- 4Self-reported symptoms are discrepant with observations of behavior
- 5Evidence from formal psychologic evaluation that the person has significantly misrepresented current status
- 1
- EBehavior meeting necessary criteria from groups B, C, and D are not fully accounted for by psychiatric, neurologic, or developmental factors
Diagnostic Categories for MPRD
- IDefinite MPRD
- 1Presence of substantial external incentive (criterion A)
- 2Definitive evidence of intent (criterion C1 or D1)
- 3Behaviors meeting the criteria for definitive intent (C1 or D1) are not fully accounted for by psychiatric, neurologic, or developmental factors (criterion E)
- 1
- IIProbable MPRD
- 1Evidence of significant external incentive (criterion A)
- 2Two or more types of probable evidence of intent from criterion B (B1–B5), criterion C (C2–C5), and/or criterion D (D2–D6). This evidence must be well-validated and have a known error rate
- 3Behavior meeting necessary criteria from groups B, C, and D are not fully accounted for by psychiatric, neurologic, or developmental factors (criterion E)
- 1
- IIIPossible MPRD
- 1Evidence of significant external incentive (criterion A)
- 2Evidence does not rise to the level sufficient for a diagnosis of probable MPRD➢Only one type of quantitative probable evidence of intent from criterion B (B1–B5), criterion C (C2–C5), and/or criterion D (D2–D6)OR➢One or more forms of qualitative evidence of intent from criterion B (B1–B5), criterion C (C2–C5), and/or criterion D (D2–D6)OR➢Evidence sufficient for a diagnosis of MPRD is present, but criterion E is not met
- 1
Appendix 3: Series of Computations to obtain the Statistical Estimate of Base Rate or Prevalence in the Full Sample
- 1All positive cases × PPP = TP cases
- 2All negative cases × NPP = TN cases
- 3All negative cases − TN cases = FN cases
- 4TP cases + FN cases = All malingering cases
- 5All malingering cases / All cases × 100 = sample BR
NOTE. BR, base rate or prevalence; FN, number of false negative cases; FP, number of false positive cases; NPP, negative predictive power; PPP, positive predictive power; TN, number of true negative cases; TP, number of true positive cases.
References
- Diagnostic and statistical manual of mental disorders.4th ed, text revision. American Psychiatric Association, Washington (DC)2000
- Antecedent probability and the efficiency of psychometric signs, patterns, or cutting scores.Psychol Bull. 1955; 52: 194-216
- Chronic pain disability exaggeration/malingering and submaximal effort research.Clin J Pain. 1999; 15: 244-274
- Pain clinic management of medico-legal litigants.Injury. 1998; 29: 305-308
- Characteristics and frequency of malingering among patients with low back pain.Pain. 1986; 25: 357-364
- Base rates of malingering and symptom exaggeration.J Clin Exp Neuropsychol. 2002; 24: 1094-1102
- Accident neurosis.Br Med J. 1961; 1: 919-925
- Accident neurosis.Br Med J. 1961; 1: 992-998
- Public attitude monitor (survey).Insurance Research Council, Oak Brook1992
- Public attitude monitor (survey).Insurance Research Council, Oak Brook1993
- The prevalence of cognitive malingering in persons reporting exposure to occupational and environmental substances.Neurotoxicology. 2006; 27: 940-950
- Assessment of malingering.in: Larrabee G.J. Forensic neurospsychology: a scientific approach. Oxford Univ Pr, New York2005: 115-158
- Assessment of malingering after mild head trauma with the Portland Digit Recognition Test.J Clin Exp Neuropsychol. 1993; 15: 170-182
- Portland Digit Recognition Test manual.2nd ed. Private Publication, Portland1993
- Test of Memory Malingering manual.MultiHealth Systems, New York1996
- Green's Word Memory Test for Window's: user's manual.Green's Publishing Inc, Edmonton2005
- Validation of malingered amnesic measures with a large clinical sample.Psychol Assess. 1994; 6: 218-224
- Wechsler adult intelligence scale.3rd ed. Psychological Corp, San Antonio1997
- The California Verbal Learning Test in the detection of incomplete effort in neuropsychological evaluation.Psychol Assess. 1995; 7: 463-471
- Assessment of response bias in mild head injury: beyond malingering tests.J Clin Exp Neuropsychol. 2001; 23: 809-828
- California verbal learning test.Research ed. Psychological Corp, New York1987
- California verbal learning test.2nd ed. Psychological Corp, New York2000
- MMPI-2: manual for administration and scoring.Univ of Minnesota Pr, Minneapolis1989
- On the diagnosis of malingered pain related disability: lessons from cognitive malingering research.Spine J. 2005; 5: 404-417
- Diagnostic criteria for malingered neurocognitive dysfunction: proposed standards for clinical practice and research.Clin Neuropsychol. 1999; 13: 545-561
- Assessment of malingered neuropsychological deficits.Oxford Univ Pr, New York2007
- Evaluation of malingered neurocognitive disorders.in: Rizzo M. Esslinger P.J. Principles and practice of behavioral neurology and neuropsychology. W. B. Saunders Co, Philadelphia2004: 1077-1089
Greve KW, Bianchini KJ. Classification accuracy of the Portland Digit Recognition Test in traumatic brain injury: results of a known-groups analysis. Clin Neuropsychol; In press.
- Classification accuracy of the test of memory malingering in traumatic brain injury: results of a known-groups analysis.J Clin Exp Neuropsychol. 2006; 28: 1176-1190
- WAIS digit span-based indicators of malingered neurocognitive dysfunction: classification accuracy in traumatic brain injury.Assessment. 2005; 12: 429-444
- Pain, malingering, and performance on the WAIS-III Processing Speed Index.J Clin Exp Neuropsychol. 2006; 28: 1218-1237
- Pain, malingering, and the WAIS-III Working Memory Index.Spine J. 2006; 6: 61-71
- Sensitivity and specificity of MMPI-2 validity scales and indicators to malingered neurocognitive dysfunction in traumatic brain injury.Clin Neuropsychol. 2006; 20: 491-512
- California verbal learning test indicators of Malingered Neurocognitive Dysfunction: sensitivity and specificity in traumatic brain injury.Assessment. 2006; 13: 46-61
- Epidemiology in medicine.Little, Brown and Co, Boston1987
- Assessment of response bias in mild head injury: beyond malingering tests.J Clin Exp Neuropsychol. 2001; 23: 809-828
- Malingering on the social security disability consultative exam: predictors and base rates.Clin Neuropsychol. 2008; 22: 529-546
- Base rates of negative response bias and malingered neurocognitive dysfunction among criminal defendants referred for neuropsychological evaluation.Clin Neuropsychol. 2007; 21: 899-916
- Malingering and uncooperativeness in psychiatric and psychological assessment: prevalence and effects in a German sample of claimants.Psychiatry Res. 2008; 157: 191-200
- Evidence-based medicine: how to practice and teach EBM.3rd ed. Elsevier Churchill Livingstone, New York2005
- Assessment of feigned cognitive impairment: a neuropsychological perspective.Guilford Pr, New York2007
- The MMPI-2 fake bad scale in detection of noncredible brain injury claims.in: Boone K.B. Assessment of feigned cognitive impairment: a neuropsychological perspective. Guilford Pr, New York2007
- An overview of the clinical use of dynamic posturography in the differential diagnosis of balance disorders.J Vestib Res. 1999; 9: 223-252
- Role of dynamic posturography (Equitest) in the identification of feigned balance disturbances.Acta Otorhinolaryngol Ital. 2004; 24: 8-12
- Posturographic performance in patients with the potential for secondary gain.Otolaryngol Head Neck Surg. 2000; 122: 11-18
- Motor dysfunction profiles in traumatic brain injury and postconcussion syndrome.J Int Neuropsychol Soc. 1996; 2: 477-485
- A new set of criteria for evaluating malingering in work-related vestibular injury.Otol Neurotol. 2005; 26: 686-690
- Effect of effort versus volume on forced expiratory flow measurement.Am Rev Respir Dis. 1988; 138: 1002-1005
- Simultaneous bilateral testing: validation of a new protocol to detect insincere effort during grip and pinch strength testing.J Hand Ther. 2002; 15: 242-250
- Classification of somatization and functional somatic symptoms in primary care.Aust N Z J Psychiatry. 2005; 39: 772-781
- Somatoform disorders among first-time referrals to a neurology service.Psychosomatics. 2005; 46: 540-548
- The psychology of spine surgery.American Psychiatric Association, Washington (DC)2003
- The use of presurgical psychological screening to predict the outcome of spine surgery.Spine J. 2001; 1: 274-282
- Psychosocial issues: their importance in predicting disability, response to treatment, and search for compensation.Neurol Clin. 1999; 17: 149-166
- A review of psychological risk factors in back and neck pain.Spine. 2000; 25: 1148-1156
- Accuracy of deception judgments.Pers Soc Psychol Rev. 2006; 10: 214-234
- Pediatric malingering: the capacity of children to fake believable deficits on neuropsychological testing.J Consult Clin Psychol. 1988; 56: 578-582
- The detection of deception.Neurol Clin. 1995; 13: 255-265
- Do first impressions count?.Eur Spine J. 1998; 7: 218-223
- Prospects for faking believable deficits on neuropsychological testing.J Consult Clin Psychol. 1978; 46: 892-900
- A validity index for the MMPI-2.Arch Clin Neuropsychol. 2002; 17: 157-169
Article Info
Footnotes
No commercial party having a direct financial interest in the results of the research supporting this article has or will confer a benefit on the authors or on any organization with which the authors are associated.
Identification
Copyright
© 2009 American Congress of Rehabilitation Medicine. Published by Elsevier Inc. All rights reserved.
