BACKGROUND: Antiepileptic drugs (AEDs) are also prescribed for therapeutic indications other than epilepsy (EPI), namely, psychiatric disorders (PSY). Our aim was to develop an algorithm able to distinguish between EPI and PSY among childbearing age women based on differences in AED exposure in these patient groups. METHODS: Two groups of women (18-45 years) with EPI or PSY treated with AEDs in the first semester of 2010 or 2011 were extracted from paper or electronic medical charts of specialized centers. Through the prescription database of Bologna Local Health Authority (Italy), AEDs, treatment schedule and co-treatments were collected for each patient. A prescription-based hierarchical classification system was developed. The algorithm obtained was subsequently validated on internal and external data. RESULTS: Eighty-one EPI and 94 PSY subjects were recruited. AED monotherapy was the most common choice in both groups (69% EPI vs 79% PSY). Some AEDs were used only in EPI, others exclusively in PSY. Co-treatments with antipsychotics (6% vs 67%), lithium (0% vs 9%), and antidepressants (7% vs 70%) were fewer in EPI than in PSY. The hierarchical classification system identified antipsychotics, SSRIs (Selective Serotonin Reuptake Inhibitors), and number of AEDs as variables to discriminate EPI and PSY, with an overall error rate estimate of 9.7% (95%CI: 5.3% to 14.1%). CONCLUSION: Among the differences between EPI and PSY, prescription data alone allowed an algorithm to be developed to diagnose each childbearing age woman receiving AEDs. This approach will be useful to stratify patients for risk estimates of AED-treated patients based on administrative databases.
BACKGROUND: Antiepileptic drugs (AEDs) are also prescribed for therapeutic indications other than epilepsy (EPI), namely, psychiatric disorders (PSY). Our aim was to develop an algorithm able to distinguish between EPI and PSY among childbearing age women based on differences in AED exposure in these patient groups. METHODS: Two groups of women (18-45 years) with EPI or PSY treated with AEDs in the first semester of 2010 or 2011 were extracted from paper or electronic medical charts of specialized centers. Through the prescription database of Bologna Local Health Authority (Italy), AEDs, treatment schedule and co-treatments were collected for each patient. A prescription-based hierarchical classification system was developed. The algorithm obtained was subsequently validated on internal and external data. RESULTS: Eighty-one EPI and 94 PSY subjects were recruited. AED monotherapy was the most common choice in both groups (69% EPI vs 79% PSY). Some AEDs were used only in EPI, others exclusively in PSY. Co-treatments with antipsychotics (6% vs 67%), lithium (0% vs 9%), and antidepressants (7% vs 70%) were fewer in EPI than in PSY. The hierarchical classification system identified antipsychotics, SSRIs (Selective Serotonin Reuptake Inhibitors), and number of AEDs as variables to discriminate EPI and PSY, with an overall error rate estimate of 9.7% (95%CI: 5.3% to 14.1%). CONCLUSION: Among the differences between EPI and PSY, prescription data alone allowed an algorithm to be developed to diagnose each childbearing age woman receiving AEDs. This approach will be useful to stratify patients for risk estimates of AED-treated patients based on administrative databases.