Identifying studies evaluating susceptibility factors for chemical health assessments: A case study focused on...

Identifying susceptibility factors for adverse health effects from chemical exposures is an important aspect of characterizing human health impacts. However, to date, an efficient approach for identifying these factors has not been established. To address this limitation, two approaches were utilized to find studies that contained susceptibility information using methylmercury (MeHg) developmental neurotoxicity (DNT) as a case study. Both approaches start with a comprehensive literature search of 5 databases on MeHg followed by keyword filtering for potential epidemiology studies; however, the approaches diverged for the subsequent steps. Approach 1 initially included screening of all 7,531 studies captured by the human filter, but was modified when it was determined that 96% of studies found to include susceptibility information were captured by a dose-response filter. Approach 2 developed a susceptibility filter to limit the screening needed. Approach 1 resulted in the identification of 172 studies with information on MeHg DNT susceptibility. Approach 2 reduced screening by 52%, but only captured 74% of PECO-relevant studies when applied to the final study set. Although Approach 2 reduced screening by 12% compared with the use of the dose-response filter in Approach 1, the decreased detection of relevant studies precludes its use in most cases. Expected technological advances that allow refinement of a susceptibility filter to improve performance would be advantageous because of the potential further reduction in screening burden. However, at this time, Approach 1, involving the application of a dose-response filter, is currently recommended for identifying epidemiology papers with information on susceptibility factors.