Abstracts

Rationale: Given the frequent occurrence of depression in persons with epilepsy, it is important to effectively detect it so that it can be adequately managed. Many depression screening tools have been validated for use in epilepsy, though an optimal tool has yet to be established. We aimed to synthesize the literature evaluating the validity of depression screening tools in epilepsy. Methods: The search strategy contained terms relating to epilepsy, depression assessment, and validity. MEDLINE, EMBASE, and PsycINFO were searched from inception until April 2016. Title and abstract review, full-text screening and data abstraction were completed in duplicate. Studies were included if they evaluated the validity of depression screening tools and reported estimates of diagnostic accuracy (e.g. sensitivity, specificity, predictive values, etc.) in persons with epilepsy. Study quality was assessed using the Quality Assessment of Diagnostic Accuracy Studies Version 2 (QUADAS-2) which includes four categories pertaining to risk of bias (patient selection, index test, reference standard, and flow and timing). Medians were calculated for the estimates of diagnostic accuracy when possible. Results: Of the 21,201 abstracts screened, 38 studies met the eligibility criteria. Validity was assessed for sixteen different screening tools and six different reference standards were used. The most commonly validated screening tools were the Neurological Disorders Depressive Inventory for Epilepsy (NDDI-E) (n= 26), the Becks Depression Inventory (n=12) and the Hospital Anxiety and Depression Scale Depression Subscale (n=10). The most commonly used reference standard was the Mini International Neuropsychiatric Interview (MINI) (n=19). The NDDI-E was most commonly validated at a cutpoint of >15, at which it had a median sensitivity (Se) of 84% and specificity (Sp) of 87%. However, the NDDI-E had the best balance of median Se and Sp at a cutpoint of >13 (Se: 84%, Sp: 86%). In terms of study quality, 13 of the 38 studies had an unclear risk of bias in at least one of the four QUADAS-2 rating system categories while seven studies had a high risk of bias in at least one of these categories. Meta-analysis was not possible due to heterogeneity (variation in cutpoints assessed and reference standards used), as well as a failure to report confidence intervals. Conclusions: Although many studies validated depression screening tools in persons with epilepsy, estimates of diagnostic accuracy were not consistently reported. In addition, the MINI was commonly used as a reference standard although it is not generally considered a gold standard tool. Cutpoints were also often selected post hoc based on the study sample, which may overestimate the validity of the scales in practice. Overall the NDDI-E was the most commonly validated tool, is publically available, free, has been validated in many languages and is easy to administer, making it feasible to use in a variety of settings. The best tool suited to a particular setting will ultimately depend on available resources. Funding: N/A