Applying systematic review methods in chemical risk assessment

Whaley, Paul (2021) Applying systematic review methods in chemical risk assessment. PhD thesis, UNSPECIFIED.

[img]
Text (2021whaleyphd)
2021whaleyphd.pdf - Published Version
Available under License Creative Commons Attribution.

Download (46MB)

Abstract

Context : Chemical risk assessment has traditionally been dependent on “narrative” approaches for synthesising evidence about potential health harms from exposure to chemical substances. However, narrative reviews are recognised as being vulnerable to a range of methodological shortcomings which introduce bias and inconsistency into the summarisation of scientific evidence. This is likely to be a contributing factor in a number of controversies about the safety of chemical substances. The potential value of systematic review methods for improving the transparency and validity of chemical risk assessments was arguably first articulated in the mid-2000s. By 2015, the first major frameworks for conducting systematic reviews of environmental health evidence had been published. What was not well understood at the time was how systematic review, as a technically exacting methodology originally developed for evaluating the effectiveness of interventions in healthcare, might be adapted to the specific workflows and evidence streams of chemical risk assessment. Objectives : The aim of this Thesis is to investigate how systematic review methods can be applied to the conduct of chemical risk assessment. This overall aim is broken down into four specific objectives: to identify practical challenges and knowledge gaps which impede the implementation of systematic review methods in chemical risk assessment; to define a consensus view on key recommended practices for the planning and conduct of systematic reviews in the environmental health sciences; to examine how “biological plausibility” as a concept fundamental to risk assessment is accommodated in systematic review methodologies; and to describe the role of ontologies in making evidence accessible for use in systematic chemical assessments. Discussion : The use of systematic review methods should improve the validity, utility and transparency of chemical risk assessments. However, the successful implementation of systematic review methods hinges on addressing a number of challenges, including the development of guidance for their conduct in environmental health contexts, and the technical development of methods where systematic review approaches need to be adapted to the specific requirements of chemical risk assessment. In terms of developing guidance, a detailed set of recommendations for the conduct of systematic reviews in environmental health and toxicological research was developed. These “COSTER” recommendations identify 70 practices across eight performance domains that will help ensure consistent and high standards for the growing number systematic reviews on environmental health topics. In terms of technical development of methods, “biological plausibility” is a concept used by risk assessors to describe the extent to which an experimental surrogate or knowledge of relevant biological mechanisms are informative of a systematic review conclusion. Through examination of 12 case examples it is concluded that “biological plausibility” is in fact already accommodated in the systematic review process under the assessment of the indirectness or external validity of evidence; however, the considerations which risk assessors take into account when assessing biological plausibility should be absorbed into the assessment of external validity of studies. Finally, examination of the concept of biological plausibility demonstrates the extreme heterogeneity and volume of data which has to be accommodated in chemical risk assessments. The role of ontologies in Knowledge Organisation Systems is examined as a key enabler of scaling up of systematic review methods to handling the volume of evidence which needs to be analysed if tens of thousands of chemicals, covering potentially millions of studies, are to be reviewed systematically.

Item Type:
Thesis (PhD)
ID Code:
150396
Deposited By:
Deposited On:
05 Jan 2021 15:20
Refereed?:
No
Published?:
Published
Last Modified:
15 Jun 2021 08:00