HERO: Health and Environmental Research Online
The Assessment Process
EPA uses risk assessment to characterize the nature and magnitude of health risks to humans (e.g., residents, workers, recreational visitors) and ecological receptors (e.g., birds, fish, wildlife) from chemical contaminants and other stressors, that may be present in the environment. Read more about risk assessment.
Systematic and Transparent
The assessment process is designed to assure the highest scientific integrity, in a transparent and iterative manner with the public. The analysis methods vary by discipline, but all risk assessment follows the same basic steps. Before beginning an assessment, a planning and scoping stage is undertaken, during which an integrated plan is developed that lays out the purposes and issues.
The HERO Database System plays a part in the assessment process in the following areas.
The first step in conducting a systematic risk assessment is the literature search. The quality of the final product is dependent on a thorough, comprehensive, unbiased literature search. To that end, an extensive collection of databases is searched.
Search strategies are developed by experienced information science professionals who work with the scientists to ensure the most complete results. New, emerging technologies are employed to conduct literature searches, ensuring broad and deep results. Using federated searching - searching across multiple databases at the same time - duplicates are eliminated and efficiency is gained. Given the geometric explosion of scientific literature in the 21st century, efficiency is becoming increasingly important.
The primary source for scientific assessments is the peer-reviewed literature. Also called "refereed," these are scholarly works that represent the latest original research in the field. The process of peer review involves rigorous screening by subject matter experts before acceptance for publication.
HERO classifies references as coming from peer-reviewed journals using Ulrich's Periodicals Directory . Consult the Related Links page for a growing list of databases of peer-reviewed literature searched by NCEA scientists.
Other Reference Types
Some topic areas have important data in the "gray" literature, which is generally defined to mean publications that do not adhere to strict bibliographic standards. These references are still valuable and consulted in the assessment development process. Technical reports from government and industry are found in this classification.
Conference proceedings are another important source for scientists in search of the current research. These publications, which may be available in print or electronic-only formats, often contain reports of research still in progress, before the researchers have published in peer-reviewed journals. They can provide insight into new directions and methods.
Every reference in HERO is identified by the reference type and source.
A critical part of the literature search process is participation from the public. EPA employs a variety of techniques to solicit additional studies: peer review workshops, public positing of literature searches and solicitation via the Federal Register and EPA website. Please participate in the process by suggesting studies for HERO.
The results of a comprehensive literature search, such as those done by NCEA scientists in health and environmental effect of pollutants and substances, can be overwhelming. The body of scientific literature is growing exponentially, far beyond human capacity to absorb and synthesize information. This explosion means that scientists spend an inordinate amount of time categorizing and developing methods to integrate conflicting and redundant evidence. This time-consuming effort could further introduce undesirable bias due to the sheer impossibility of the task.
Visual interpretation of citations is a cutting-edge area the HERO Database System brings to the assessment process. Through text mining, the information discovery process illuminates new relationships and gems heretofore overlooked, enabling the highest degree of confidence in data and selection of key studies. This is most obvious in the areas of epidemiology and computational biology, where weeding through the mass of published literature can be frustrating. These huge datasets are made more manageable and understandable in short shrift.
A series of modules using advanced algorithms for training and deploying different models in the various disciplines, enable rapid, comprehensive retrieval and organization. Data verification and analysis techniques draw on sophisticated data management, which, when added to full-text retrieval and semantic processing, assure the highest scientific integrity in data quality.
Recent advances in the area of automatic information extraction (IE) from unstructured sources has opened up new avenues for querying, organizing and analyzing scientific data culled from the world of biomedical and environmental research. The field of IE has its genesis in natural language processing, using computer-aided intelligence and computational linguistics to extract information in a comprehensive, unbiased manner. The techniques used can then be applied to retrieve vast amounts of unstructured text from the published literature for an optimal synthesis system.