David Teira

Frequentist versus Bayesian Clinical Trials

Publisher Summary This chapter presents an overview of the philosophical debate on frequentist versus Bayesian clinical trials. The comparison between these approaches has focused on two main dimensions: the epistemology of the statistical tools and the ethics of the different features in each experimental design. The philosophical debate on the flaws of frequentist randomized clinical trials (RCTs) was elaborated presenting a number of arguments that hold independently of any conception of probability. The conflicts more directly related to the frequentist foundations of RCTs: namely, the ethical issues involved in randomization and in the stopping rules that may close a trial before it reaches the targeted sample size, are also explained. An overview of the regulatory uses of RCTs, discussing the main alternatives considered for drug testing in different countries is presented. The necessity of a middle ground between the potential flexibility of Bayesian approaches and the necessity of standardized Bayesian procedures is emphasized.

Facts, norms and expected utility functions

In this article we explore an argumentative pattern that provides a normative justification for expected utility functions grounded on empirical evidence, showing how it worked in three different episodes of their development. The argument claims that we should prudentially maximize our expected utility since this is the criterion effectively applied by those who are considered wisest in making risky choices (be it gamblers or businessmen). Yet, to justify the adoption of this rule, it should be proven that this is empirically true: i.e. that a given function allows us to predict the choices of that particular class of agents. We show how expected utility functions were introduced and contested in accordance with this pattern in the 18th century and how it recurred in the 1950s when Allais made his case against the neo-Bernoullians.

Causality, Impartiality and Evidence-Based Policy

The overall aims of this chapter are to compare the use of randomised evaluations in medicine and economics and to assess their ability to provide impartial evidence about causal claims. We will argue that there are no good reasons to regard randomisation as a sine qua non for good evidential practice in either science. However, in medicine, but not in development economics, randomisation can provide impartiality from the point of view of regulatory agencies. The intuition is that if the available evidence leaves room for uncertainty about the effects of an intervention (such as a new drug), a regulator should make sure that such uncertainty cannot be exploited by some party’s private interest. We will argue that randomisation plays an important role in this context. By contrast, in the field evaluations that have recently become popular in development economics, subjects have incentives to act strategically against the research protocol which undermines their use as neutral arbiter between conflicting parties.

Disease-mongering through clinical trials

Our goal in this paper is to articulate a precise concept of at least a certain kind of disease-mongering, showing how pharmaceutical marketing can commercially exploit certain diseases when their best definition is given through the success of a treatment in a clinical trial. We distinguish two types of disease-mongering according to the way they exploit the definition of the trial population for marketing purposes. We argue that behind these two forms of disease-mongering there are two well-known problems in the statistical methodology of clinical trials (the reference class problem and the distinction between statistical and clinical significance). Overcoming them is far from simple.

On the impartiality of early British clinical trials

Did the impartiality of clinical trials play any role in their acceptance as regulatory standards for the safety and efficacy of drugs? According to the standard account of early British trials in the 1930s and 1940s, their impartiality was just rhetorical: the public demanded fair tests and statistical devices such as randomization created an appearance of neutrality. In fact, the design of the experiment was difficult to understand and the British authorities took advantage of it to promote their own particular interests. I claim that this account is based on a poorly defined concept of experimental fairness (derived from T. Porters ideas). I present an alternative approach in which a test would be impartial if it incorporates warrants of non-manipulability. With this concept, I reconstruct the history of British trials showing that they were indeed fair and this fairness played a role in their acceptance as regulatory yardsticks.

Blinding and the Non-interference Assumption in Medical and Social Trials

This paper discusses the so-called non-interference assumption (NIA) grounding causal inference in trials in both medicine and the social sciences. It states that for each participant in the experiment, the value of the potential outcome depends only upon whether she or he gets the treatment. Drawing on methodological discussion in clinical trials and laboratory experiments in economics, I defend the necessity of partial forms of blinding as a warrant of the NIA, to control the participants’ expectations and their strategic interactions with the experimenter.

A Contractarian Solution to the Experimenter’s Regress

Debiasing procedures are experimental methods aimed at correcting errors arising from the cognitive biases of the experimenter. We discuss two of these methods, the predesignation rule and randomization, showing to what extent they are open to the experimenter’s regress: there is no metarule to prove that, after implementing the procedure, the experimental data are actually free from biases. We claim that, from a contractarian perspective, these procedures are nonetheless defensible since they provide a warrant of the impartiality of the experiment: we only need proof that the result has not been intentionally manipulated for prima facie acceptance.

Philosophy of Medicine: Causality, Evidence and Explanation

in particular. Since Hemmo and Shenker do not mention the role of time-translation invariance, they do not seem to address the proposal satisfactorily and so fail in their rebuttal. Chapter 9 is devoted to the notion of measurement, while in chapter 10, Hemmo and Shenker discuss the problem of retrodiction (the fact that retrodicting the past, contrarily to predicting the future, is inconsistent with our memories). Chapter 11 deals with the Gibbsian formulation of statistical mechanics, of which the authors give a novel interpretation. In chapter 12, Hemmo and Shenker address the notion of information and the thermodynamic of computation, while in the final chapter, the authors arrive at Maxwell’s demon, which they consider the prototype counterexample to the laws of thermodynamics, and explain how it is consistent with statistical mechanics. In addition, there are two appendices: appendix A provides an example of a ‘Demonic’ evolution, whereas appendix B considers some issues related to the role of probabilities in quantum mechanics. This second appendix is divided into three parts: in B1, the authors address David Albert’s proposal connected to the Ghirardi– Rimini–Weber theory; in B2, they question the role of typicality in Bohmian mechanics; and in B3, they argue that Maxwell’s demon is compatible with quantum mechanical dynamics with or without collapse. To conclude and summarize, I think that the account that Hemmo and Shenker propose may be compelling only to those who believe that Boltzmann’s explanation does not work, something of which they have not yet managed to convince me. Be that as it may, the book provides an interesting, original, and valued contribution to the literature in the foundations of statistical mechanics that surely deserves attention and careful reading.

Debiasing Methods and the Acceptability of Experimental Outcomes

Why do scientists reach an agreement on new experimental methods when there are conflicts of interest about the evidence they yield? I argue that debiasing methods play a crucial role in this consensus, providing a warrant about the impartiality of the outcome regarding the preferences of different parties involved in the experiment. From a contractarian perspective, I contend that an epistemic pre-requisite for scientists to agree on an experimental method is that it be neutral regarding their competing interests. I present two medical experiments (on smallpox inoculation and mesmerism) in which debiasing procedures such as blinding and data tabulation provided warrants of impartiality that allowed people to agree on an experimental design even if they disagreed on the outcome.

Quasi- and Field Experiment

Field trials and quasi-experiments are comparative tests in which we assess the effects of one intervention (or a set thereof) on a group of subjects as compared to another intervention on another group of similar characteristics. The main difference between field trials and quasi-experiments is in the way the interventions are assigned to the groups: in the former the allocation is randomized whereas in the latter it is not. We are going to see first the different roles played by randomization in medical experiments. Then we will discuss how controlled field trials, originating in psychology, spread to the social sciences throughout the twentieth century. Finally, we will show how the idea of a quasi-experiment appeared around a debate on what constitutes a valid test and what sort of controls guarantee it.