Low-calorie sweeteners and health outcomes: A demonstration of rapid evidence mapping (rEM). (February 2019)

• Record Type: Journal Article
• Authors: Lam, Juleen; Howard, Brian E.; Thayer, Kristina; Shah, Ruchir R.
• Is Part Of: Environment international. Volume 123(2019)
• Page Start: 451
• Page End: 458
• Publication Date: 2019-02
• Subjects: rEM rapid evidence mapping -- LCS low calorie sweeteners -- PECO Participants, Exposure, Comparator, and Outcomes -- GRAS generally recognized as safe -- MeSH Medical Subject Headings -- SRDR The Systematic Review Data Repository; Rapid evidence mapping -- Evidence map -- Systematic review -- Evidence-based methodology -- Low calorie sweeteners -- Artificial sweeteners; Environmental protection -- Periodicals; Environmental health -- Periodicals; Environmental monitoring -- Periodicals; Environmental Monitoring -- Periodicals; Environnement -- Protection -- Périodiques; Hygiène du milieu -- Périodiques; Environnement -- Surveillance -- Périodiques; Environmental health; Environmental monitoring; Environmental protection; Periodicals; 333.705
• DOI: 10.1016/j.envint.2018.11.070
• Languages: English
• ISSNs: 0160-4120

Abstract

Abstract: Background: "Evidence Mapping" is an emerging tool that is increasingly being used to systematically identify, review, organize, quantify, and summarize the literature. It can be used as an effective method for identifying well-studied topic areas relevant to a broad research question along with any important literature gaps. However, because the procedure can be significantly resource-intensive, approaches that can increase the speed and reproducibility of evidence mapping are in great demand. Methods: We propose an alternative process called "rapid Evidence Mapping" (rEM) to map the scientific evidence in a time-efficient manner, while still utilizing rigorous, transparent and explicit methodological approaches. To illustrate its application, we have conducted a proof-of-concept case study on the topic of low-calorie sweeteners (LCS) with respect to human dietary exposures and health outcomes. During this process, we developed and made publicly available our study protocol, established a PECO (Participants, Exposure, Comparator, and Outcomes) statement, searched the literature, screened titles and abstracts to identify potentially relevant studies, and applied semi-automated machine learning approaches to tag and categorize the included articles. We created various visualizations including bubble plots and frequency tables to map the evidence and research gaps according to comparison type, population baseline health status, outcome group, and study sample size. We compared our results with a traditional evidence mapping of the same topic published in 2016 (Wang et al., 2016). Results: We conducted an rEM of LCS, for which we identified 8122 records from a PubMed search (January 1, 1946–May 1, 2014) and then utilized machine learning (SWIFT-Active Screener) to prioritize relevant records. After screening 2267 (28%) of the total set of titles and abstracts to achieve 95% estimated recall, we ultimately included 297 relevant studies. Overall, our findings corroborated those of Wang et al. (2016) and identified that most studies were acute or short-term in healthy individuals, and studied the outcomes of appetite, energy sensing and body weight. We also identified a lack of studies assessing appetite and dietary intake related outcomes in people with diabetes. The rEM approach required approximately 100 person-hours conducted over 7 calendar months. Conclusion: Rapid Evidence Mapping is an expeditious approach based on rigorous methodology that can be used to quickly summarize the available body of evidence relevant to a research question, identify gaps in the literature to inform future research, and contextualize the design of a systematic review within the broader scientific literature, significantly reducing human effort while yielding results comparable to those from traditional methods. The potential time savings of this approach in comparison to the traditional evidence mapping process make it a potentially powerful tool for rapidly translating knowledge to inform science-based decision-making. Highlights: rEM is a protocol driven and expeditious approach to evidence mapping. rEMs can identify research gaps and areas of opportunity for systematic review. Use of machine learning tools significantly reduce human effort. The rEM case study took 100 person-hours, offering potential time savings. rEMs offer similar findings compared to a traditional evidence mapping.