Fig. 1. Illustration of electronic prioritization approaches.

1 A. Schematic illustration of electronic prioritization of literature depicting references clustered by similarity using natural language processing. 1 B. Illustration depicting clusters containing relevant seed references (circled blue clusters). Clusters were ranked by the number of seed studies included. 1C. Visualization of identified clusters. Clusters were organized into groups (A–F) on the basis of the number of approaches that identified the cluster such that Group A contains clusters harboring seed references identified by six approaches and Group F contains clusters harboring seed references identified by a single approach. All references in the top four groups (A–D) were manually screened for inclusion based on PECO criteria. Low scoring groups (E, F) were subjected to additional machine learning approaches to capture relevant references for manual screening. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)