Popperian epidemiology and the logic of bi-conditional modus tollens arguments for refutational analysis of randomised controlled trials.

Abstract

Popperian epidemiology is a biomedical science tool based on the hypothesis-deductive method and the falsifiability of scientific hypotheses. This article explores the applicability of the refutationist logic tools in the analysis of a randomised controlled trial (RCT), the randomised Aldactone evaluation study (RALES). This was carried out by using bi-conditional modus-tollens arguments of the type (i) P-then-Q(n) and (ii) Q(n)-If-X(P), X(P) being a set of potential falsifiers of Q(n) as part of the explicit falsity-content of P. In this model, P is the main hypothesis and Q(n) one or more logical predictions to be tested. The X(P) argument represents inclusion criteria, exclusion criteria and conditional criteria of the RCT so every P-then-X(P) argument should be fulfilled in canonical form to corroborate P-then-Q(n). Thus, falsifiability of a RCT would be determined by the empirical content of the conditional argument Q(n)-If-X(P) and its external validity would be determined by the empirical content of X(P). In this way it would be possible to mathematically assess the external validity of a RCT if the observational predicates of the X(P) argument in a given population are known. According to this popperian model, applicability of the RCT results to clinical practice implies transferring of all its empirical content, in other words, the totality of its truth and falsity contents. Thus, to ignore the explicit falsity-content of a RCT such as RALES may jeopardise its potential benefits in clinical practice as suggested by recent studies.

• Authors: Elard Koch, Alvaro Otarola, Tomas Romero, Aida Kirschbaum, Esteban Ortuzar
• Journal: Medical hypotheses (Med Hypotheses) Vol. 67 Issue 4 Pg. 980-8 ( 2006) ISSN: 0306-9877 [Print] United States
• PMID: 16730919 (Publication Type: Journal Article, Research Support, Non-U.S. Gov't)

Topics

• Clinical Trials as Topic
• Epidemiologic Methods
• Humans
• Logic
• Models, Biological
• Randomized Controlled Trials as Topic (standards)
• Research Design (standards)