Use this URL to cite or link to this record in EThOS:
Title: Evidence evaluation and the epistemology of causality in medicine
Author: Auker-Howlett, Daniel
ISNI:       0000 0004 9351 734X
Awarding Body: University of Kent
Current Institution: University of Kent
Date of Award: 2020
Availability of Full Text:
Access from EThOS:
Access from Institution:
Establishing causal claims is important across medicine. For example, a new drug is effective at lowering blood pressure if, and only if, the drug is a cause of a decrease in blood pressure. To work out if the drug does indeed cause a decrease in blood pressure requires a methodology for causal evaluation. The dominant methodology for causal evaluation in medicine is that of Evidence Based Medicine (EBM). A competing methodology is EBM+. The aim of this thesis is to defend EBM+ as a more complete epistemology of causality in medicine than that of EBM. An epistemology of causality in medicine must answer two questions: i) what kinds of evidence should be used to evaluate causal claims?; ii) how should that evidence be evaluated? On the one hand, EBM and EBM+ differ on i), as EBM holds that only evidence from clinical studies can be used to evaluate causation, whereas EBM+ admits evidence from both clinical and mechanistic studies, which utilise the methods of the biomedical sciences. On the other hand, they mostly agree on ii), as they agree that evidence should be evaluated explicitly and transparently. Where they do disagree on ii), they do so only because they disagree on what kinds of evidence should be evaluated. What counts as evidence for causality is thus the key difference between the two methodologies. The plan of the thesis is as follows. In Chapter 1 I introduce both EBM and EBM+. I then set out the conceptual framework I use, as well as introduce the case studies I analyse, throughout the thesis. I defend an evidentially pluralist position on what counts as evidence for causation in Chapter 2. EBM+ is evidentially pluralist and so is provisionally a closer approximation to a complete epistemology of causality in medicine than EBM. It is not enough to just be evidentially pluralist when it comes to evaluating causality, as EBM+ must also be able to provide methods for the evaluation of evidence of mechanism, particularly that obtained from mechanistic studies. In the remainder of the thesis I analyse an evaluative framework, built on EBM+ principles, namely, 'Evaluating evidence of mechanisms in medicine' (EEMM) (Parkkinen et al., 2018b). In general, one can have feasibility, practical, conceptual and malleability worries about any methodology of causal evaluation. To allay feasibility worries, I use EEMM to carry out a systematic review of evidence of mechanism obtained from mechanistic studies in Chapter 3. I then defend EEMM against some practical worries about establishing mechanisms on the basis of mechanistic studies in Chapter 4. A conceptual worry about the evaluative process in EEMM concerns how to characterise judgements about the strength of evidence. In particular, in Chapter 5 I defend the interpretation used in EEMM of 'quality of evidence', which is put in terms of 'stability of confidence'. Malleability worries concern the role of expert judgement in evidence evaluation: whether subjective choices influence judgements to an inappropriate extent. This worry is motivated by the medical nihilism thesis, which claims that we should only ever hold low confidence in the effectiveness of medical interventions. As medical nihilism poses a problem to any methodology for causal evaluation in medicine, I reject it in Chapter 6. I do however concede that the need to make subjective choices during evaluations makes the malleability of methods a worry that any evaluative framework must contend with. Accordingly, in Chapter 7 I consider the potential to use formal models of belief to constrain the influence of subjectivity on judgements. I then use lessons learnt in Chapters 5 and 7 to ameliorate a potential example of malleability in EEMM, by formulating additional guidance in Chapter 8. I motivate further extensions to the EEMM framework in Chapter 9. Such extensions involve integrating it with a clinical study evaluator, the Grading of Recommendations, Assessment, Development and Evaluation (GRADE) framework. I finish by summarising the claims of this thesis, drawing consequences for real world practice, and identifying open questions. I conclude that EBM+ as a methodology of causal evaluation is feasible, conceptually defensible, and can counter malleability concerns. Moreover, where appropriate I propose recommendations on how to improve on the EBM+ analysis and the EEMM methodology. Parts of this thesis are derived from, or were developed during the production of, previously published articles. A version of the introduction to EEMM found in §3.2 can be found in Abdin et al. (2019, Section 4), to which I was the sole contributor of that section. In Chapter 2, I appeal to, rather than alter and include, ideas found in papers to which I was a contributor: Where relevant, I cite Auker-Howlett and Wilde (2019) and Auker-Howlett and Wilde (2020), as the research and analysis of the case study found in this article was carried out in collaboration with Michael Wilde.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID:  DOI: Not available