Personality assessment inventory sex differences in people with epileptic and functional seizures

Abstract

Objective

Previous research has shown that the Personality Assessment Inventory (PAI) is useful in differentiating individuals with functional seizures (FS) from those with epileptic seizures (ES). In these two groups, sex differences in PAI validity and clinical subscales have not been investigated. Elucidating sex differences could improve the understanding of how males and females experience FS or ES and may improve differential diagnosis and treatment planning.

Method

We examined sex differences across PAI scales, subscales, and validity profiles amongst adults with FS (n = 62) and ES (n = 55). Participants were evaluated and classified at the Johns Hopkins Hospital Epilepsy Monitoring Unit based on continuous video electroencephalography (cEEG) confirmed diagnoses. Among valid profiles, we conducted the following analyses: calculation of odds ratios (OR) and associated confidence intervals (CI) for each clinical, treatment, and interpersonal scale; logistic regression models to examine whether diagnostic status could be predicted by PAI scale elevations and sex; positive predictive values (PPV) and negative predictive values (NPV) to understand the diagnostic utility of scale and subscale elevations; and receiver operating characteristic (ROC) curves to determine the area under the curve (AUC) and identify PAI scales with the greatest capacity to distinguish FS presence/absence.

Results

There were no significant differences in validity scales between FS and ES groups, both in the combined sample and when stratifying by sex (p’s > 0.05). On a clinical scale level, FS females produced higher elevations on stress (STR) and somatic complaints (SOM) (OR = 4.17, OR = 6.95, p’s < 0.05), and FS males reported higher anxiety (ANX), anxiety-related disorders (ARD), and non-support (NON) (OR = 8.0, OR = 11.25, OR = 15.2, p’s < 0.05) than sex-matched ES patients. On a subscale level, FS females were more likely than ES females to have clinically elevated conversion (SOM-C) and somatization (SOM-S) scales (OR = 4.72, OR = 5.78, p’s < 0.05), and FS males were more likely than ES males to report clinically elevated SOM-C, physiological anxiety (ANX-P), and trauma-related distress (ARD-T) (OR = 10.53, OR = 9.52, OR = 9.52, p’s < 0.05). ROC findings revealed that SOM-C (AUC = 0.72) and ANX-P (AUC = 0.78) had the greatest AUCs for FS females and males, respectively. In the combined sex sample, FS patients were more likely than ES patients to endorse mildly elevated paranoia (PAR) and STR and clinically elevated SOM, depression (DEP), and ARD (ORs = 4.86, 3.52, 5.10, 3.01, 4.38, respectively, p’s < 0.05); when entering these scales in a logistic regression model, SOM was the best predictor of FS. However, overall classification rates did not exceed 80 % (PPV = 75.5 %, NPV = 65.5 %). Among subscales, the combined-sex FS cohort was more likely to endorse elevations on SOM-C, SOM-S, ANX-P, and ARD-T than the ES cohort (ORs = 5.24, 3.82, 4.38, 4.38, respectively, p’s < 0.05); when entering these scales in a logistic regression model, SOM-C was the best predictor of FS. Within the combined sex sample, SOM-C and STR had the highest AUCs (0.70 and 0.68, respectively).

Conclusions

Our findings indicate that males and females with FS are both more likely than those with ES to endorse higher conversion symptoms but otherwise have different characteristics. These differences might influence predisposition to the development of FS or perpetuation of symptoms and could be useful in determining optimal treatment approaches.