2776

While the process for developing search strings for physical-chemical properties is relatively straight-forward due to availability of existing data presented in a controlled manner, identifying published fate and human environmental chemical exposure-related studies is difficult. Challenges also exist for identifying health-related studies, but controlled vocabularies for indexing and PubMed Medical Subject Headings (MeSH) help to identify such studies, while relatively fewer developed vocabularies exist for identifying fate and exposure studies. In addition, the scope of ‘exposure’ is broad and comprises several sub-categories. Provision of these search strings will facilitate and standardize the approach for systematic review of scientific evidence needed for transparent chemical assessment. To develop the search strings, a team of topic-specific experts and librarians was assembled and developed keyword lists. To create the physical-chemical string, librarians used the keyword list and a series of iterative exchanges with the expert to draft, and revise strings based on performance. For fate strings, librarians drafted strings utilizing the keyword list and analysis tools such as the Keyword Analysis Tool (KAT) and SWIFT-Review’s fingerprinting feature, and on-topic reference lists to determine performance; exposure strings will be developed in a similar manner. The quality of all test strings will be further evaluated by human manual screening of on- and off-topic search string generated results. This presentation overviews various approaches taken to develop the fate, chemical-physical property and exposure literature search strings, including development and testing of broad verses minimal search strings (where we identify terms that are unambiguous and/or necessary to retrieve the greatest number of on-topic references while minimizing retrieval of off-topic references); discusses the organization of keywords and search strings into main and subcategory headings; and outlines the process for incorporating keyword analysis tools. The recall and precision performance results of the broad versus minimal search strings are also presented.