Abstract
Introduction
ADHD has long been understood to be a disorder of brain development and leads to symptoms of inattention, hyperactivity, and impulsivity (Castellanos et al., 2002). Many studies have shown abnormal brain activity and even structure in ADHD populations compared with normal controls (Jacobson et al., 2018; Shaw et al., 2007). Although ADHD is considered a neurodevelopmental condition, symptoms are known to persist into adulthood. Although full criteria from the Diagnostic and Statistical Manual of Mental Disorders (5th ed.; DSM-5; American Psychiatric Association, 2013) may not continue to be met, research indicates that up to 50% of individuals diagnosed as a child or adolescent may continue to symptomatic well into adulthood (Biederman, Petty, Evans, Small, & Faraone, 2010; Faraone, Biederman, & Mick, 2006).
Actual rates of ADHD persistence into adulthood, as defined by the DSM-5, can be difficult to ascertain for many reasons. Many adults continue to be symptomatic but learn compensatory behaviors to decrease functional impact (Harrison & Edwards, 2010). In addition, research has shown the potential for secondary gain as many adults seek the assistance of stimulant medication and/or academic accommodations in postsecondary schooling that often accompanies an ADHD diagnosis (Clemow & Walker, 2014; Sansone & Sansone, 2011; Sollman, Ranseen, & Berry, 2010).
As a result, many institutions recommend neuropsychological (NP) testing to document or systematically measure attentional or executive impairment. NP testing can be helpful in determining attention or executive/working memory impairment, though there are no specific tests or test battery that are “sensitive or specific to serve as diagnostic indices” for the condition (American Psychiatric Association, 2013). Also, attention and executive/working memory impairment can occur for many different reasons and are present in many different psychiatric conditions, and testing demonstrating impairment does not necessarily equate an ADHD diagnosis.
Performance Validity Testing (PVT) has become a standard in the field of NP evaluation to ensure adequate effort during testing. Such has been routinely applied to standard NP testing for several decades now to help ensure accurate inferences are made from test data. Although failure to reach recommended PVT cut-offs on various tests indicates suboptimal effort, they do not indicate why. Often, inadequate effort is interpreted as volitional and an attempt to malinger symptoms yet PVT failure can occur for many reasons, such as distraction from headache or physical discomfort, fatigue, confusion or frustration with medical procedures, unfamiliarity with testing processes, and many others (Heilbronner et al., 2009).
Recently, there have been a number of studies that demonstrate failed PVT rates in adults presenting for ADHD evaluation. Rates have been variable, depending on measures and methods used, as well as criteria which defines “failed performance.” Although technically one failed PVT can indicate poor effort, more than one failed test is recommended to make this inference (Heilbronner et al., 2009; Proto et al., 2014). Although failure can occur for reasons other than intentional feigning of symptoms (above), failing should not be considered due to the deficiencies in the construct being measured. Several studies have demonstrated those diagnosed with ADHD are able to routinely pass PVT (Jasinski et al., 2011; Sollman et al., 2010; Williamson et al., 2014). In addition, those having severe brain injuries with clinical problems with attention have also been shown to pass PVTs (Macciocchi, Seel, Alderson, & Godsall, 2006).
On the lower end of sample base rates for failed PVT in adults presenting for ADHD evaluation, Pella, Hill, Shelton, Elliot, and Gouvier (2012) showed that in a large sample of community college students, the base rates for probable malingering was just above 10%. In a sample of 268 adults referred for ADHD assessment, Marshall and colleagues (2010) measured probable insufficient effort using a compilation of embedded and stand-along PVTs, along with discrepancies between observed and reported ADHD symptoms. They found that approximately 22% of their sample engaged in probable feigning of symptoms.
On the higher end of sample base rates, Suhr, Hammers, Dobbins-Buckland, Zimak, and Hughes (2008) found a 31% base failure rate of PVT in 85 young adults (mean age = 22.7) at a university setting referred for ADHD evaluation. Of note, the criteria used to define “fail” on PVT was rather liberal, with a below cut-off score on any one of the four subtests of the Word Memory Test (WMT), a well-established PVT, used as criteria. When two or more subtests were used, this figure dropped to 25%. In total, 11% of their sample failed all four subtests, indicating a likelihood of not only poor effort on testing but also possible volitional intent to score poorly.
In a study examining 66 psycho-educational assessments for ADHD and learning disorders (LD), Sullivan, May, and Galbally (2007) found that the LD evaluations only had a 22.4% failure rate on a commonly used PVT (WMT). However, compared with LD evaluations, those who were being evaluated for ADHD had a near 50% failure rate (47.6%). The comparison between the two evaluations is interesting given the differences in potential benefits the two diagnoses can provide. Although a diagnosis of an LD provides the potential for academic accommodations, a diagnosis of ADHD provides the potential for both academic accommodations and stimulant medication. Such differences in potential for secondary gain is important to consider and was interpreted in their study as the main mediator between the differences in PVT failure rates between the LD and ADHD groups.
Although several studies have demonstrated base rates of PVT failure in college populations presenting for ADHD evaluation, to date there has only been one study examining this issue in U.S. military populations (Shura, Denning, Miskey, & Rowland, 2017). Although their approach primarily examined subjective report of symptoms (symptom validity testing), they did include one PVT measure and found a near 20% failure rate (19.3). Focusing more heavily on accurate symptom report, they found a 45% (44.7) failure rate on a commonly used personality test (MMPI-2-RF [Minnesota Multiphasic Personality Inventory–2 Restructured Form]) with embedded symptom validity measures. However, as delineated in other work, though symptom and performance validity can be related, they are not identical constructs (Heilbronner et al., 2009) and assessing adequate cognitive effort despite concerns of inaccurate symptom report is vital for ADHD evaluations. In addition, as mentioned previously, more than one PVT administration is recommended to determine adequate cognitive effort in NP testing (Heilbronner et al., 2009; Proto et al., 2014).
The purpose of this study is to expound on the work done by Shura and colleagues and further examine PVT failure rates in ADHD evaluations in military populations. To date, there are no known studies that have examined this issue within active duty (AD) military populations. In addition to incentives already discussed, military populations carry additional potential motivators for needing stimulant medication, given that working odd and long hours is common and the need for continued alertness and vigilance remains high despite this. In addition, although this can fluctuate (depending on ever-changing military needs), the use of stimulant medication can even hinder or preclude worldwide deployment. Such is a desirable outcome for those who wish to avoid overseas assignment, separation from family, or service in hazardous areas of the world. Although there are no known differences between AD military members and veterans that would dramatically effect ADHD outcomes, this study seeks to further work done by Shura and colleagues by using additional PVTs to test for cognitive effort according to clinical guidelines established in the literature.
The main research questions of this study are as follows:
Equally valuable is to examine the characteristics of PVT failure rates and their association with other important variables that should be captured in ADHD evaluations. Given that DSM-5 criteria has specifically designated childhood history of ADHD (if not explicitly diagnosed than at least some indication of problems from an early age), ruling out comorbid conditions, and establishing at least some degree of impaired functioning is needed for the diagnosis, this study sought to examine PVT failure and its relationship with these variables. Finally, as an effort to further explore PVT failures in ADHD evaluations, the sociodemographic variables of age and estimated premorbid functioning were examined and their relationship with PVT performance.
Method
Participants
Data were extracted from 51 psychological reports from individuals who were referred for an ADHD evaluation from November 2016 to May 2018. All evaluations were conducted by a single doctoral-level, fellowship-trained neuropsychologist (author), and data were extracted from all referrals during the above time interval (none excluded). All referrals came from other medical providers (clinical social workers, psychiatrists, or primary care physicians or physician assistants) with the main presenting complain as impairment of attention. All participants were on AD status in the U.S. military at the time of evaluation and stationed (permanently or temporarily) at Fort Hood, Texas. Participant ages ranged from 19 to 43, with an average age being approximately 29 years old (28.55), relatively older than many samples taken from university campuses. Approximately 69% were male (n = 35). All were actively engaged in their respective fields, none were going through a Medical Evaluation Board for retirement, separating from the Army, or otherwise not gainfully engaged in their respective military roles. In total, 51% of the sample identified as Caucasian, 27% African American, 20% Hispanic, and 2% as Asian. If on a prescribed stimulant medication at the time of the evaluation, they did not take it on the day of testing (specifically asked at time of testing).
Other information taken from reports of interest for this study were (a) estimated premorbid intellectual functioning, (b) prior history of ADHD as child or adolescent, (c) history of other mental health condition, and (d) report and evidence of some occupational or other impairment at the time of the evaluation.
An estimate of premorbid intellectual functioning was taken from scores on the Armed Services Vocational Aptitude Battery (ASVAB). The ASVAB is a multiple-choice test used by the U.S. Military entrance Processing Center to determine entrance eligibility and occupational placement. The test is a compilation of various achievement tests which measure arithmetic reasoning, general science knowledge, verbal reading, and comprehension skills and other abilities and skills. It is designed similar to conventional intellectual tests in that verbal and nonverbal abilities are heavily weighted. In fact, the ASVAB has long been used as an accurate predictor of premorbid intelligence, and correlation with traditional IQ tests have been quite high (Orme, Brehm, & Ree, 2001). Overall General Technical (GT) score was used, which is a compilation of one’s performance on Word Knowledge, Paragraph Comprehension, Arithmetic Reasoning, and Mechanical Comprehension scores. Scaling of test is similar to traditional intelligence tests (Wechsler Scales of Intelligence) and has a mean of 100 and standard deviation of 15. Very low scores (below 80) are excluded in that applying candidates with such scores do not meet military entrance criteria. This likely contributed to the sample having a mean score roughly a half standard deviation above national average intelligence scores for adults (107.1, SD = 12.9).
For estimated premorbid intelligence, even with the shifted distribution the overall sample appeared evenly distributed among ability-types. Although not equal intervals, if groups are divided into GT scores of <99, 100 to 114, and >115, sample groupings were 18, 17, and 16, respectively. Participant characteristics are summarized in Table 1 below.
Sample Characteristics.
Note. ASVAB GT = Armed Services Vocational Aptitude Battery General Technical (score).
Childhood history of ADHD was assessed based on reported history and coded. As is well-known in the field, obtaining history of childhood ADHD can be difficult due to poor recall of childhood history. Lack of diagnosis, stimulant medication use, or academic accommodations is not proof that condition did not exist, as many individuals come from backgrounds which lack resources to adequately assess and treat ADHD. Parent resistance or negative attitudes toward ADHD diagnosis and treatment may also negatively affect history of care. Due to this, and unavailability of childhood academic records or collaborative proxy information, history of ADHD was coded in a manner that captured these varying levels of historical report of symptoms or diagnosis. Specifically, participants were assigned to three groups: (a) those that reported a bona fide diagnosis with treatment (medication, therapy or academic accommodations), (b) those who reported a “suspected” condition but no diagnosis or treatment, (i.e., many complaints and some indication of problems), and (c) those who credibly denied any history of attention impairment as a child. Although academic records prior to military training were not used, training records after military enrollment were used. Military training for occupational placement (Advanced Individual Training, or “AIT”), often occurs in a classroom setting, and can significantly resemble formal education settings. Such training records were used in determining groups (specifically Group 2; those that did not report diagnosis as a child but where AIT records clearly indicated problems with academic performance).
Participates were also coded for either having history of or no history of mental health condition other than ADHD. This was coded not solely based on personal report, but also following a review of medical records. For participants referred with no accompanying mental health symptoms or history, attention problems were interpreted as stand-alone problems, and not in the context of other adjustment, anxiety, depressive symptoms, or other mental health concerns. Groupings for this variable were as follows: (0) no other mental health history, (1) history of anxiety-related condition, (2) adjustment disorder, (3) major depressive disorder or dysthymia, or (4) other mental health condition. Because of the low number in groupings, for purposes of analyses groups of individuals with other mental health history were all combined. Final variable groupings were bivariate, having either no or some other mental health history other than ADHD.
The last variable coded was report of impaired functioning at the time of the evaluation. As above, this variable was not a simple subjective report of impairment but also associated with at least some evidence of impaired functioning. This evidence typically came from participant military records and performance ratings and was available at the time of the evaluation. Other information that served as evidence of impaired functioning was failed or in struggling status while enrolled in military training at the time of the evaluation (to be distinguished from history of problems as described above). Information from participant spouses, commanding officer, or supervisor was also used. Objective report of impaired decision-making or impulsive behavior, such as poor financial planning, impulsive purchasing, excessive traffic violations or accidents, and so on, was also used to determine group assignments.
The majority of referrals had either an actual childhood ADHD diagnosis with medication prescription, or suspected diagnosis and reported symptoms from an early age (67%, n = 34), though the remainder of the referrals, roughly 33% (n = 17), had no history of ADHD diagnosis or reported symptoms as a child or adolescent. The majority of the sample had at least some other mental health history (65%); however, those that reported and were found to have at least some functional impairment at the time of the evaluation was nearly split even between the two groups (26 reported some functional impairment vs. 25 reported no impairment). Such characteristics are summarized below (see Table 2).
Additional Sample Characteristics.
Procedures
Research procedures were reviewed and approved by the institutional review board at Carl R. Darnall Army Medical Center in Fort Hood, Texas. As part of each evaluation, participants were seen for an hour-long intake session where all intake and descriptive information was recorded. Along with the personal variables recorded above, participants were given a variety of cognitive and psychological tests to assess for ADHD. Overall, a flexible battery was used, and although some evaluations included additional tests not included in this study (personality tests), those that are included were all given in the same order. Details of the tests used are given below. Although personality tests were given they were not the focus of the study and were given variably and are not used in the analysis.
To measure PVT pass/fail group differences in age and estimated premorbid intellectual functioning level, independent sample t tests were conducted with age and premorbid intellectual functioning as dependent variables. Assumptions of the tests were met and results are reported as equal variances assumed. Given that all other analyses included comparing actual versus estimated group proportions, chi-square tests for independence were conducted.
Outcome Measures
The first and standard PVT given to all participants was the Victoria Symptom Validity Test (VSVT), a computer-administered PVT that has successfully demonstrated reliability and validity as a tool that accurately measures cognitive effort (Slick, Hopp, Strauss, & Thompson, 1997). Typically, a tool that is used to assess for cognitive dysfunction due to head trauma or injury, the VSVT has also shown to be effective when assessing for cognitive effort in ADHD evaluations (Frazier, Frazier, Busch, Kerwood, & Demaree, 2008). The test is a forced-choice recognition measure consisting of 24 “easy” and “difficult” items (Slick et al., 1997). A cut score of 19 on “hard” items was considered PVT “failure” and follows recommended cut scores by Frazier and colleagues when using the test in ADHD evaluations (Frazier, Naugle, Busch, Haggerty, & Youngstrom, 2007).
As is the case with all cognitive testing measures, there are various reasons for failed performance, and to avoid false positives a second PVT was given in the case of VSVT failure. The additional test for all participants was the Test of Memory Malingering (TOMM; Tombaugh, 1996). As with the VSVT, the TOMM is a forced-choice recognition test and considered a well-validated and commonly used NP test to assess for cognitive effort during NP testing (Tombaugh, 1996). Recommended cut-off score of <45 on Trial 2 was used as PVT failure. Unlike the VSVT, the TOMM is a test administered face-to-face, which can potentially rule-out false positives that may stem from failed performance because of computer-administration. In the rare events that participants failed initial VSVT administration but passed TOMM administration, their performance on an actual test of attention (and built-in validity measures to be discussed below) as well as behavioral observations were used to make the final determination of whether participants met study criteria of PVT “pass” or “fail.” Those that failed initial VSVT administration, but passed the TOMM and whose behavioral observations gave indication of adequate effort (i.e., not responding to phone, paying attention to instructions, adequately engaged during testing) were given a “pass” distinction on PVT and coded accordingly.
Finally, participants completed a well-researched, standardized measure of attention performance (Conners’ Continuous Performance Test, 2nd Edition [CPT-II]; Conners, 2000). The performance profile provides 11 subtest scores for inattention, impulsivity, and vigilance, and an overall confidence index that indicates whether performance better matches a “clinical” versus “nonclinical” profile. Because of the inherent problems of reducing performance to a single score, participant performance was categorized as “impaired” if at least six of the 11 subtests were outside normal limits or if the overall confidence index better matched a clinical versus nonclinical profile, as defined by 75% confidence or higher. Categorized in this manner, the overall sample performed relatively equally on the CPT-II, with 23 participants (45%) performing within normal limits (55% performing poorly).
Results
In reference to the primary research question of this study, PVT failure rates were similar to those found in civilian populations. Overall, 19 of the 51 participants (37.3%) were considered PVT failures, according to the criteria delineated in the “Method” section above. Although not an exhaustive comparison, this appears to fall within ranges of studies conducted on university campuses (see “Introduction” section), though on the upper-end of the range. Although not in line with the procedures of this study, overall PVT failure rate would have increased to 41% if only initial PVT score was used as measure of cognitive effort (21 out of 51 failed using this criteria).
Men were more likely than women to be considered PVT failures, though this was not a statistically significant finding (43 vs. 25%, χ2 = 1.5, df = 1, p = .22). Younger age and lower estimated premorbid intelligence were associated with PVT failure (T = 4.09, p < .01 and T = 2.31, p = .02, respectively), though these variables were mildly related (r = .31). Results are summarized in Table 3.
Results of Gender, Age, and Estimated IQ on PVT Groups.
Note. PVT = Performance Validity Test; ASVAB GT = Armed Services Vocational Aptitude Battery General Technical (score).
Those who did and did not fail PVT did not differ on the other variables of interest. Individuals who had a bona fide childhood history of ADHD, those with a suspected history only, and those with no history equally passed or failed PVTs (χ2 = 2.5, df = 2, p = .28). Collapsing those with definite or suspected history and comparing that group against those with no reported history of condition or suspicion of condition also yielded similar results (χ2 = 2.06, df = 1, p = .15). With respect to those who had no other mental health history other than ADHD, six participants failed PVTs (6 / 18 = 33%), compared with 13 failures in the “other mental health history” group (13 / 33 = 39%) which was not a significant finding (χ2 = .74, df = 1, p = .39). There was virtually no difference in PVT failure classification in those that had some degree of functional impairment versus those that did not. The groups themselves were nearly equally divided (51% reported problems, 49% no problems), and each group had nine and 10 failures, respectively.
Not surprisingly, PVT pass/fail categorization was significantly different in groups who performed within normal limits or poorly on the CPT-II, with the PVT failure group performing poorly more often than the PVT nonfailure group (χ2 = 5.45, df = 1, p = .02).
Discussion
As demonstrated in the “Introduction” section, PVT failure rates of those seeking ADHD evaluations outside of military contexts fall in the approximate range of 10% to 46%. The results of this study show that PVT failure rates for ADHD evaluations are similar in U.S. military populations, though toward the higher end of the spectrum. More notable is that this study used fairly stringent criteria for PVT failure description. Using more liberal criteria that have been used in other studies, such as a single PVT failure, would have resulted in rates closer to those found by Sullivan and colleagues (2007) (41%).
Compared with other research, the results obtained in this study fall in the more frequent rate of PVT failure while using relatively stringent criteria. This indeed continues to substantiate that poor cognitive effort is a concern for adults referred for ADHD evaluation. In addition, in agreement with Shura and colleagues (2017), this study indicates that poor effort is a concern for those seeking ADHD evaluations in military populations. This study extended the findings from Shura and colleagues and indicates poor cognitive effort occurs not only in those with a history of military service but also in those that are currently serving. Reasons for this are likely similar to those found in civilian populations, mainly the possibility of academic accommodations and/or stimulant medication use, but may also be due to avoid worldwide deployment. Stimulant medication may also be increasingly tempting due to the highly frequency of sleep-deprivation typically found in military populations, and occupational demands of high vigilance.
The PVT failure group was overall younger and had lower estimated premorbid IQs than nonfailures. Although it may be considered natural to conclude that lower estimated premorbid intelligence is linked with greater PVT failure because of diminished cognitive capacity, there are several reasons why this is an unlikely accurate interpretation of the data. As mentioned in the “Procedures” section discussing the ASVAB, military candidates whose score falls in the low range (approximately GT score of <80) are denied military entrance and therefore were naturally excluded from this study. Even if this were not the case, PVT literature has demonstrated no differences in performance between controls and individuals of low IQ (Full Scale Intelligence Quotient [FSIQ] < 70) when properly motivated (Green & Flaro, 2015). Even individuals who have sustained severe brain injury and have serious subsequent cognitive deficits have been shown to routinely be able to pass PVTs (Macciocchi et al., 2006). Such are interesting findings, and although it is unknown why such group differences exist, it is consistent with research that has demonstrated higher levels of education is associated with decreased rates of PVT failure (Webb, Batchelor, Meares, Taylor, & Marsh, 2012).
There were no group differences between those who did and did not pass PVT and groups differing on childhood history of ADHD, other mental health condition, or current functional impairment. Given that the premise behind an evaluation for ADHD is to present evidence of the condition, for those with no childhood history or evidence of impaired functioning, it may follow that evaluees may be aware that a manifestation of cognitive impairment would help their cause. Therefore, it was somewhat surprising that there were no significant findings in these groups.
There are several strengths and limitations of this study. This article, as well as many others, have adequately highlighted possibilities for secondary gain in ADHD evaluations for adults. Not only is there a need to ensure adequate cognitive effort is given during evaluation, but due to problems of subjective report, more methods need to be taken to ensure accuracy of reported history and report of symptoms. This was a strength for this study, given that medical, occupational performance, and academic records were easily available. This is not always the case, and additional studies in military populations with access to this information can be a great resource to ADHD evaluation literature. Another strength of this study was that it was conducted in an AD military population. This expounded on other research with military veterans and demonstrates that findings with veteran populations can be generalized to those on AD. Significant findings in yet another population with different demands and culture, as other study populations, reinforce the need for cognitive effort testing in ADHD evaluations.
Limitations in this study include the relatively small sample used. Additional studies with more participants would be useful to further examine the issue of cognitive effort in ADHD evaluations within military samples. In addition, another limitation was the research methods used. Although the study was meant for exploratory purposes, given that the main research hypothesis was supported further work can be done to elucidate factors associated with those that do and do not give sufficient effort when being evaluated for ADHD. This is not so much done to identify those who give poor effort but rather to increase the probability those with true impairment are accurately identified and receive the help they need. This project focused on group differences only; stronger research methods can not only identify factors associated with poor or adequate effort during ADHD testing but also test predictive models that may assist clinicians in properly assessing for the condition.
In summary, in addition to the points discussed above, this study continues to reinforce the need for cognitive effort testing in ADHD evaluations and extends this need to populations other than just in university settings. In addition to the main research question of this study, this study showed an increased need for vigilance in proper screening individuals referred for ADHD evaluations. Although the presence of childhood ADHD is often difficult to ascertain, this study showed that a patient report history or problems consistent with ADHD was absent in approximately 30% of all referrals. It also showed that a large proportion (33%, n = 17) of referrals fell in the above average range of estimated intellectual functioning (ASVAB GT score ≥ 115). Although attentional difficulties are certainly not constrained to only average and below IQ functioning, providing accommodations to those with accelerated abilities is controversial. In fact, as has been argued by prominent researchers in the field, “it is inherently unfair to grant advantages to those who are actually not subnormal” (Barkley, Murphy, & Fischer, 2008). The American With Disabilities Act (ADA) also favors this interpretation, and views proper interpretation of “impairment” in the context of ADHD as those whose impairment is defined by below average performance compared with population norms and not relative to a narrow, highly specialized, and accomplished subset of individuals whose estimate of cognitive ability actually falls above average (Barkley et al., 2008).
In connection with this, the need for demonstrated impaired functioning, another requirement for the diagnosis (per DSM-5), was also absent in nearly half of all referrals. Although the evaluation criteria in this study were likely more rigid than what is typical, and more information available in a military setting (performance records easily accessible), the results indicate vigilance in the referral process, as well as the evaluative process, is needed for accurate assessment of ADHD. If the interpretation of impairment as favored by prominent researchers in the field and the ADA are accepted, those with evidence of above average intelligence or high occupational or educational accomplishment may not be appropriate candidates for testing and evaluation.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
