Validity testing in veterans with epileptic seizures and psychogenic nonepileptic seizures

Identification of clinically meaningful subgroups among patients with psychogenic nonepileptic seizures (PNES) or epileptic seizures (ES) is of potential value for assessing prognosis and predicting therapeutic response. Invalid performance on validity tests has been associated with noncredible complaints and worse cognitive test scores, and may be one such classification criteria. We studied invalid performance in Veterans with PNES or ES, and the association of invalid performance with cognitive test scores and subjective complaints.

Patients were consecutive admissions to three veterans affairs (VA) epilepsy monitoring units. Evaluations included two validity tests: the Test of Memory Malingering (TOMM); and the Minnesota Multiphasic Personality Inventory-2-Restructured Form (MMPI-2-RF) over-reporting validity scales. We compared the frequency of invalid performance on the TOMM or MMPI-2-RF in patients diagnosed with PNES vs. ES. We evaluated the association of invalid performance with scores on the Repeatable Battery for Assessment of Neuropsychological Status (RBANS), and four subjective symptom measures including the Beck Depression Inventory-II, and Quality of Life in Epilepsy-31.

Invalid TOMM performance was found in 25.3% of Veterans diagnosed with PNES and 10.8% of those with ES (p = .03). Invalid reporting on the MMPI-2-RF was found in 35.9% of the PNES group vs. 15.3% of the ES group (p = .01). Effects of valid vs. invalid reporting on external measures were similar for ES and PNES groups. Patients with invalid vs. valid TOMM performance had lower scores on the RBANS (p < .001). Patients with invalid performance had greater complaints on all subjective measures, with largest effect sizes for the MMPI-2-RF validity scales (p < .001).

In Veterans admitted for evaluation of poorly controlled seizures, invalid performance on validity tests was not uncommon. Cognitive test results and subjective reports from patients with invalid performance may not be credible. These observations have implications for the analysis of clinical trials, where primary and secondary outcomes often rely on self-report measures.