You are here:
Advancing Kinetic Modeling for Polychlorinated Biphenyl (PCB) Mixtures Relevant to Human Exposures: A Case Study Applying the KeywordAnalysis Tool to PCBs Congener-Specific Half-Lives
Citation:
Shipkowski, K., J. Trgovcich, N. Vetter, B. Ingle, R. Silva, L. Carlson, D. Kapraun, P. Schlosser, AND G. Lehmann. Advancing Kinetic Modeling for Polychlorinated Biphenyl (PCB) Mixtures Relevant to Human Exposures: A Case Study Applying the KeywordAnalysis Tool to PCBs Congener-Specific Half-Lives. International Society for Environmental Epidemiology, Washington DC, NC, August 23 - 27, 2020.
Impact/Purpose:
KAT is a software tool designed to increase efficiency in literature screening; it allows users to calculate metrics on the utility of each keyword. By applying keywords related to half-life, elimination, and clearance, KAT identified 5,956 potentially relevant studies from the PCB literature database. A text analytics approach was implemented to further select studies for data extraction.
Description:
Toxicity resulting from human exposure to PCBs is highly dependent on the absorption, distribution, metabolism, and excretion (ADME) properties of individual PCB congeners. A database of half-life of elimination values for individual congeners was assembled to advance the development of a comprehensive kinetic model for PCB mixtures relevant to human exposures. A novel strategy was employed to identify relevant literature. First, a literature search was conducted using keywords related to PCBs; this generated a database of 58,915 studies. Next, the Keyword Analysis Tool (KAT) was applied to select studies related to PCB congener-specific half-lives across mammalian species. KAT is a software tool designed to increase efficiency in literature screening; it allows users to calculate metrics on the utility of each keyword. By applying keywords related to half-life, elimination, and clearance, KAT identified 5,956 potentially relevant studies from the PCB literature database. A text analytics approach was implemented to further select studies for data extraction. Specifically, sentences from the full text of each study that harbored a keyword were reviewed in conjunction with titles and abstracts to ascertain study relevance. This effort identified 312 studies, 98 of which contained half-life values for 95 PCB congeners across mammalian species. Data from 33 studies identified half-life values for 36 PCB congeners in humans across lifestages, biological matrices, and exposure routes. A retrospective analysis of the KAT search results showed that of 64 keywords used, only 28 were associated with relevant studies, and only 24 keywords would have captured all relevant studies. Additionally, by including keywords relevant to ecological studies, KAT successfully helped prioritize other studies more likely to be relevant to PCB elimination in mammalian species. Thus, KAT identified a highly effective keyword set that can be applied to increase efficiency of future literature searches for chemical half-lives of PCB elimination in humans.