SIRS-2 Not as Useful as SIRS for Identifying Feigned Psychopathology

Anthony M. Tarescavage, Patton State Hospital and Kent State University
David M. Glassmire, Patton State Hospital

The Structured Interview of Reported Symptoms (SIRS) underwent a major revision in 2010 yielding the SIRS-2. The new test has since been criticized for several potential problems, particularly in terms of its sensitivity to feigned psychopathology. For this reason, the purpose of this study was to examine the concordance between SIRS and SIRS-2 classifications and sensitivity estimates in an archival sample of 263 criminal defendants (215 males, 48 females) who were admitted to a high-security state psychiatric hospital for restoration of competency to stand trial. In a subgroup of 39 presumed feigning patients who elevated 1 or more collateral measures of feigning (primarily the M-FAST) at conservative cutoffs, we found marked discrepancies between the sensitivity of the SIRS (.87) and SIRS-2 (.54). The marked differences in sensitivity were partially explained by a global interpretation discordance rate of 47%, with discordance primarily resulting from SIRS-based feigning cases being classified as indeterminate on the SIRS-2. Follow-up analyses of intercorrelations and percentile distributions indicated that the new SIRS-2 scales may lack utility in the assessment of feigning because of problems relating to the construct validity of the scales and their interpretive cutoffs. Future directions in research and clinical practice are discussed, with added emphasis on the significant limitations of archival pretrial forensic samples for identifying presumed genuine groups necessary to calculate specificity estimates (which were meaningfully higher for the SIRS-2 in this sample). Overall, the primary clinical implication is that feigning should remain a strong consideration in SIRS-2 cases yielding an indeterminate classification.

SIRS, SIRS-2, feigning, malingering, M-FAST

“The Structured Interview of Reported Symptoms is a structured interview to detect feigned psychopathology that assesses a variety of response styles typically used by individuals fabricating or exaggerating psychological symptoms (e.g. reporting extreme or unusual symptoms, inconsistencies between reported and observed behavior, etc.). A test taker can be classified as genuine responding, indeterminate, probable feigning, or definite feigning on each scale, and an overall SIRS classification is based on the number of scales in the probable or definite feigning range (p.488).”

“Rogers, Sewell, and Gillard (2010) have created a revised version of the test: the SIRS-2. The SIRS-2 test items and Primary Scales remain the same, but the major changes were the addition of three new measures used to classify global interpretations and, relatedly, a new decision-tree model of interpretation. The revision was prompted by a SIRS study at the Timberlawn Trauma Program in Dallas, Texas, where the authors found relatively high false-positive rates for SIRS classified cases among poly traumatized inpatients predominately presenting with dissociative identity disorder or other dissociative symptoms. Two classification measures were added specifically to reduce false-positive classification errors among trauma-exposed patients: The Rare Symptoms (RS) Total scale and the Modified Total (MT) Index. The new scales and their corresponding decision-tree steps were also intended for use with nontrauma-exposed samples. The researchers “added two other measures to the Supplementary Scale (SS) Index and Improbable Failure (IF) scale. The SS Index is calculated by summing four of the SIRS-2 Supplementary Scales, which include IF, Direct Appraisal of Honesty (DA), Defensive Symptoms (DS), and Overly Specified Symptoms (OS). The Primary goal of the SS Index was to identify ‘too-good-to-be-true’ SIRS profiles produced by likely feigners (p.489).”

“Despite the strengths of the SIRS-2 in terms of approved specificity as well as the more practical modified total index, several concerns have been raised since the release of the test. The main criticisms enumerated have been for the following reasons: (a) calculating classification accuracy estimates in a non-standard manner; (b) failure to provide adequate documentation in the test manual; and (c) using a development sample that is not generalizable to common forensic settings (490).”

“The purpose of the current study was to address several gaps in the literature that limit its use in forensic settings. We sought to provide intercorrelations between the SIRS and SIRS-2 scales, as well as add to the classification concordance results using an archival and ecologically valid forensic inpatient sample with a higher local base rate of feigning (i.e. predominately feigning referrals instead of consecutive inpatient admissions). We believe these estimates may be more generalizable to clinical practice, where the SIRS/SIRS-2 is most likely to be administered when clinical judgment or screening measures indicate the need for further evaluation. We expected that the use of a sample with a higher local base rate of feigning would result in relatively worse concordance estimates because Green et al. (2013) found that the poorest classification convergence was for SIRS-based indeterminate and feigning classifications (p.491).”

“The sample included 263 nonconsecutive criminal defendants (215 males, 48 females) who were admitted to a high-security state psychiatric hospital for restoration of competency to stand trial. All patients in the sample were referred for psychological evaluations by their treatment teams and subsequently administered the SIRS as part of this assessment at the discretion of the evaluator” (p.492). We “identified a presumed feigning group from the subsample of 95 patients who were concurrently administered the MMPI-2, PAI, or MFAST (in addition to the SIRS) and were specifically referred for feigning assessment. The current sample was deemed inadequate for deriving a presumed genuine responding group. Nevertheless, we did identify a presumed genuine responding group in the interest of completeness. The sample was derived from the subsample of 120 patients who were referred for a psychological evaluation and concurrently administered the MMPI-2, PAI, or MFAST (in addition to the SIRS) (p.494).”

“We found that SIRS and SIRS-2 classifications converged in only 54% of cases and the SIRS-2 had appreciably worse sensitivity estimates than the original SIRS. Although the SIRS-2 had higher specificity estimates, the presumed genuine group used to calculate these estimates in this particular sample is deemed inadequate for reasons enumerated below.” Consistent with hypotheses and the results reported by prior researchers, “we found that the SIRS-2 was associated with a meaningfully lower sensitivity estimate than the SIRS. Using standard calculations, the SIRS correctly identified 34 of 39 of the presumed feigning patients (sensitivity = .87), whereas the sirs-2 correctly identified 21 of the presumed feigning patients (sensitivity = .54). These findings indicate hat the discordance between SIRS and SIRS-2 classifications is associated with marked decrements to sensitivity (in addition to any potential improvements to specificity) (p.498).”

“The results of this study have implications for future research on the SIRS-2. Given the substantial sensitivity losses found in the current study and by other researchers, independent research does not support that the SIRS-2 is useful for identifying feigned psychopathology as the SIRS. More thorough examinations of the RS total scale are needed at this juncture, both in terms of its construct validity with other established measures of feigning as well as its functioning at alternative cutoffs. Of similar interest, it may be useful to compare the functioning of the MT Index to an alternative that includes all of the SIRS Primary Scale scores. The results of this study also point to the need for investigation of alternative MT Index cutoffs (p.501).”

“Forensic practitioners who use the SIRS-2 are likely to find that a meaningful proportion of test takers who are classified as indeterminate-evaluate or indeterminate-general by the test would have been classified as feigning by the SIRS. Because these differences partially reflect the reduced sensitivity of the SIRS-2, practitioners should not rule out the presence of feigning based on a SIRS-2 indeterminate classification. Rather, the results of the current study suggest that a large proportion of feigners will be classified as indeterminate-evaluate or, to a lesser extent, indeterminate-general. For these reasons, practitioners may need to shift their conceptualization of SIRS-based feigning assessment with the introduction of the SIRS-2, such that feigning should remain a strong consideration in cases yielding an indeterminate-evaluate classification. More broadly, these findings combined with those of prior researchers indicate it may be premature to replace the SIRS with the SIRS-2 (p.500).”

As always, please join the discussion below if you have thoughts or comments to add!