Abstracts

Rationale: Depression is a potentially life-threatening disorder that affects approximately one third of patients with epilepsy. Self-report questionnaires, such as the Personality Assessment Inventory (PAI), have been used to enhance detection of depression and other common psychiatric disorders in clinical populations. The PAI is a standardized self-report questionnaire that is designed and normed for use with psychiatric populations. It includes 11 clinical scales, including depression, which are further broken down into cognitive, affective, and physiological subscales. Endorsement of physiological or cognitive symptoms (e.g., fatigue, sleep disturbance, impaired cognition) caused by epilepsy or antiepileptic medications may confound prevalence estimates by inflating the rate of depression. The goal of the present study was to determine whether or not the PAI depression subscales (cognitive, affective, and physiological) accurately detect depression, as diagnosed by a psychiatrist, in patients with medically-intractable epilepsy. Methods: Fifty-one consecutive medically-intractable epilepsy patients underwent neuropsychological and psychiatric evaluations as part of a comprehensive evaluation for surgery candidacy. PAI depression subscale scores were compared among these patients as well as the normative and clinical populations used in the development of the PAI. Subsequent comparisons and logistic regression analyses were performed between psychiatrist-diagnosed non-depressed (n = 30) and depressed (n = 21) groups on individual PAI depression subscale items. Results: Patients with epilepsy reported significantly more affective symptoms than the PAI standardization sample, but less than the PAI clinical sample (p<.001). However, patient scores on cognitive and physiological subscales were similar to the PAI clinical sample, and significantly greater than the PAI standardization sample. Individual-item comparisons using Mann U-Whitney non-parametric analysis found that depressed patients reported a greater severity of affective and cognitive symptoms. Physiological item differences were found only on those questions pertaining to sleep. Subsequent logistic regression analyses found that the affective subscale best differentiated the patients identified as depressed or non-depressed. Conclusions: The use of self-report measures such as the PAI in patients with epilepsy is confounded by the physiological and cognitive effects of epilepsy and treatment with antiepileptic medications. As expected, self-reported cognitive and physiological symptoms failed to discriminate patients with epilepsy alone from those with co-morbid depression. In contrast, the affective subscale of the PAI successfully discriminated patients with epilepsy alone from those with co-morbid depression. Use of the PAI total depression score may lead to overestimating depression in patients with epilepsy and should therefore only be interpreted with a focus on the affective subscale.