Development of TRAEC

Initial Tool 

In the past three years, the co-Initiators have discussed and formed a consensus on the content of the initial TRAEC strategy version.

Step 1: Problem formulation

Researchers can select target environmental chemical referring to their interest and experience, and determine a general range of health effects according to their respective research fields. If the assessment is conducted for application, this process can be skipped. A systematic literature search was conducted to find evidence of the health effects of the target chemical in the target range. Researchers are recommended to prepare a simple protocol first, incorporating a list of retrieval platforms (searching engines and public databases), key words, and inclusion/exclusion criteria. Both primary research and secondary research are acceptable.

A scientific question is formulated at the beginning of the risk assessment to guide the entire research process. For assessments with a specific research purpose, the target chemicals and health outcomes have already been determined. For those without a specific research purpose, a research question should be posed in which the target health outcome is the only undetermined component. Of all possible health effects, the mechanism of most interest to the investigator is selected as the target mechanism based on the study results.

Step 2: Evidence collection

Risk assessment of environmental chemicals requires a comprehensive evaluation of available research evidence, including targeted evidence searches and the investigator's own experiments. Specific information on three aspects of epidemiological, in vivo and in vitro studies, is obtained by integrating data from different sources, such as databases, authorities, and researchers' laboratories.

 Epidemiological studies.Risk management is conducted for human health, and risk assessment is incomplete without data from human health studies. For epidemiological studies, the target population must first be identified. The investigator then needs to assess the exposure to the target chemical, such as information route of exposure, dose, etc. In addition to this, baseline information on the study population such as personally identifiable information, lifestyle, disease history, treatment and fertility is included. This process of assessing exposure is optional and in some cases may also utilize retrospective survey methods such as completing standardized questionnaires online or by telephone. In addition, to identify and quantify adverse health effects in populations associated with the target chemical, researchers can investigate the incidence under the corresponding exposure through the literature, public databases. If not found, researchers conduct their studies to collect epidemiological data to assess dose-response/effect relationships in the population.

 In vitro studies. Researchers need to consult the literature to review the relevant results of published in vitro studies. In addition, for unarticulated mechanisms, researchers can conduct their own in vitro studies to validate and supplement the literature evidence, including molecular initiation events, cellular responses, or cellular phenotypes of target chemicals, etc.

 In vivo studies. Similar to the design of in vitro studies, for mechanisms that have not been elucidated by existing in vivo studies. Researchers can conduct experiments to validate and complement organ responses and the organism's response to exposure to target chemicals.

Step 3: Evidence scoring & Conclusion

Risk assessment of environmental chemicals requires an integrated evaluation of existing research evidence, which ranges from epidemiological studies to in vivo and in vitro studies. For individual researchers, it remains elusive how to evaluate comprehensive evidence in a systematic, transparent, and consistent manner to conclude with high reliability. Therefore, the Comprehensive Evidence Scoring Rules were developed to meet such challenges.

With data from diverse sources, such as databases, authoritative institutions, and researchers’ labs. The numerical framework can be used to assess health risk by integrating four key items, including reliability scores, correlation scores, outcome fitness scores, and integrity Scores. The scores of diverse evidences are combined and the final result is calculated according to the formula to represent the overall validity of the evidence, the direction, and the strength of the association.

All evidence is recommended to be scored independently by two researchers, with their average as the final score of each item. If a score differs by 2 or more points, it needs to be arbitrated by another advanced researcher. In particular, if a study is mixed by any two or more evidence types, i.e., epidemiological/in vivo/in vitro evidence, researchers need to sum up the scores of both/all types.

Reliability Scores. The reliability scores are used to measure the quality of experimental design and implementation and here we score the study based on three components, including the reliability of epidemiological/in vivo /in vitro study design and implementation.1 point for each criterion.

1. Reliability of epidemiological study design and implementation (0-5)

The overall study design is appropriate and effective for addressing the study objectives.

Baseline data for the included and excluded populations, high and low exposure populations are balanced, comparable, and corrected for important confounding factors.

Data sources for exposure and outcomes are reliable (e.g. detection data sets, surveillance data sets, clinical diagnosis).

Descriptions that reflect the quality control in the implementation process are included, such as on-site supervision, quality control of samples or questionnaires, etc.

Endpoints of epidemiological studies are defined by clinical diagnosis or laboratory tests.

2. Reliability of in vivo study design and implementation (0-5)

The species and strains of experimental animals meet the requirements of the experimental purpose.

A reasonable exposure mode is used, with a clear explanation of duration, dosage, and routes of exposure, etc.

Intervention/rescue experiments are used to prove the association.

Endpoints of in vivo experiments involve phenotypic or pathological changes.

Endpoints of in vivo experiments involve cellular or molecular changes.

3. Reliability of in vitro study design and implementation (0-5)

The type of cells and culture method meet the requirements of the experimental purpose.

A reasonable exposure mode is used, with a clear explanation of duration, dosage, and routes of exposure, etc.

Intervention/rescue experiments are used to prove the association.

Endpoints of in vitro experiments involve cellular changes.

Endpoints of in vitro experiments involve molecular changes.

Correlation Scores. The correlation scores here reflect the direction of the association between exposure and outcome as derived from the population/in vivo/in vitro experiments described above.

1 point: a positive correlation.

0 point: no significant correlation.

-1 point: a negative correlation.

Outcome Fitness Scores. The outcome fitness scores refer to the extent to which the results of the study support the association.

Outcome fitness of the epidemiological study (1-2)

1 point: Low risk between exposure and outcome (e.g., low odds ratio, relative risk, hazard ratio, etc.).

2 points: High risk between exposure and outcome (e.g., high odds ratio, relative risk, hazard ratio, etc.).

Outcome fitness of the in vivo study (1-2)

1 point: Appearance of only molecular changes.

2 points: Appearance of both phenotype and molecular changes.

Outcome fitness of the in vitro study (1-2)

1 point: Appearance of only molecular changes.

2 points: Appearance of both phenotype and molecular changes.

Integrity Scores. Integrity scores refers to the comprehensive degree of various types of evidence. Complete evidence should include epidemiological, in vivo, and in vitro experiments. Include only one item, multiply the final score by one-third, include two items by two-thirds, and all items by 100%.

Comprehensive evidence score=

∑(Reliability Score * Correlation Score * Outcome Fitness Score)/Evidence Number * Integrety Score

Reliability Score: the reliability of study design and implementation.

Correlation Score: the direction of association between exposure and outcome obtained through the study.

Outcome Fitness Score: the extent to which the study results support the overall hypothesis.

Evidence Number: the total number of evidence involved in scoring.

Integrity Score: the comprehensive degree of various types of evidence.

Comprehensive Evidence Score: evidence strength and correlation direction.

Tool Version 1.0

The main concern of the system review is the neglect of different sources of evidence concentration. In the Tool Version 1.0, we hope to find a method that can reasonably calculate the weights of concentrations, and thus generate new versions.