Methods and technologies for improving feasibility and accuracy of preclinical systematic reviews of in vivo pain neurobiology data
File(s)
Author(s)
Soliman, Nadia
Type
Thesis
Abstract
The failure to translate preclinical success to the clinic has led researchers to question the predictive validity of animal models with inadequate study design, lack of rigour and opaque reporting potentially responsible for translational failures. Preclinical systematic reviews aim to answer a range of research questions to improve research value and translational ability of results. However, there are several challenges which limit their feasibility including the exponentially increasing, large volume of literature and resource intensive methods.
The aims were to (i) address key areas of neurobiology to provide empirical evidence to improve and guide future preclinical research and (ii) evaluate methods to improve the feasibility, efficiency, and accuracy of preclinical systematic reviews, including crowd science, machine learning for study selection, and semi-automation of data extraction (the application Graph2Data).
I performed a narrative review supported by a systematic search and study selection to explore the role of VGF in neuropathic pain. The “systematic review and mete-analysis of studies in which cannabinoids were tested for antinociceptive effects in animal models of pathological or injury-related persistent pain” quantified anti-nociceptive efficacy and assessed the reliability of the data. Crowd science and machine learning were employed and assessed. Graph2Data was further developed and integrated into the online Systematic Review Facility (SyRF).
The reviews demonstrate that improved rigour and transparency of design, conduct, analysis, and reporting is required and the difference between the animal and clinical population highlights the importance for development of better validated animal models. Increasing biological variation will likely improve the generalisability of results and bridge the gap between animal models and the clinical population. Crowd science and machine learning are beneficial, mutually supportive, viable strategies for improving feasibility. The continued development and incorporation of machine assisted methods in open-source, online systematic review software is necessary for accurate and timely evidence synthesis.
The aims were to (i) address key areas of neurobiology to provide empirical evidence to improve and guide future preclinical research and (ii) evaluate methods to improve the feasibility, efficiency, and accuracy of preclinical systematic reviews, including crowd science, machine learning for study selection, and semi-automation of data extraction (the application Graph2Data).
I performed a narrative review supported by a systematic search and study selection to explore the role of VGF in neuropathic pain. The “systematic review and mete-analysis of studies in which cannabinoids were tested for antinociceptive effects in animal models of pathological or injury-related persistent pain” quantified anti-nociceptive efficacy and assessed the reliability of the data. Crowd science and machine learning were employed and assessed. Graph2Data was further developed and integrated into the online Systematic Review Facility (SyRF).
The reviews demonstrate that improved rigour and transparency of design, conduct, analysis, and reporting is required and the difference between the animal and clinical population highlights the importance for development of better validated animal models. Increasing biological variation will likely improve the generalisability of results and bridge the gap between animal models and the clinical population. Crowd science and machine learning are beneficial, mutually supportive, viable strategies for improving feasibility. The continued development and incorporation of machine assisted methods in open-source, online systematic review software is necessary for accurate and timely evidence synthesis.
Version
Open Access
Date Issued
2021-07
Date Awarded
2021-11
Copyright Statement
Creative Commons Attribution NonCommercial Licence
License URL
Advisor
Rice, Andrew
Sponsor
Biotechnology and Biological Sciences Research Council (Great Britain)
Grant Number
BB/M011178/1
Publisher Department
Faculty of Medicine
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)