A substantial minority of our patients became spell-free with communication of the diagnosis the only intervention. Previous psychiatric diagnoses, social security payments, and gender were important predictors of outcome. Most patients stopped using emergency services, irrespective of whether or not spells continued. Outcomes other than spell frequency may be important in patients with psychogenic nonepileptic attacks.

The clinical predictors that we found provide further evidence of heterogeneity of psychogenic nonepileptic attacks populations related to antecedent factors. Some predictive clusters may have clinical utility in the early stages of assessment: in particular the combination of medically unexplained symptoms and self-harm should raise the possibility of undisclosed sexual abuse.

The data suggest a distinct subgroup of patients with late onset psychogenic nonepileptic attacks, in whom psychological trauma related to poor physical health plays a prominent role.

Diagnostic error is not entirely avoidable and a degree of uncertainty, and perforce risk, is intrinsic to the diagnostic process of epilepsy. The risks of a false negative diagnosis of epilepsy must be appraised against the also significant risks of a false positive diagnosis.

Our samples showed few significant gender differences, suggesting that other male and female populations of patients with PNES are likely to be similar also. Some of the differences we found may give insight into causation of PNES.