DISCUSSION PAPER Resistance is not futile, but neither is it always justified Kirstin Borgerson PhD Assistant Professor, Department of Philosophy, Dalhousie University, Halifax, NS, Canada Keywords clinical practice, compellingness, Mark Tonelli, research relevance, scientific validity, social value Correspondence Dr Kirstin Borgerson Department of Philosophy Dalhousie University PO Box 15000, NS Canada BRH4R2 E-mail: kirstin.borgerson@dal.ca Commentary prepared for the JECP Special Issue on Philosophy and Medicine 2013 Accepted for publication: 18 March 2013 doi:10.1111/jep.12057 Physicians do not always adjust their treatment recommendations in response to the latest research evidence, even when the research in question is judged to be methodologically rigorous. In his recent paper, ‘Compellingness: assessing the practical rel- evance of clinical research results’, Mark Tonelli identifies 12 features of clinical research that help to explain this resistance [1]. These features, which determine how compelling research results are to individual clinicians, are grouped into three catego- ries: (1) epistemic factors; (2) fit with individualized care and patient values; and (3) considerations related to the stewardship of health care resources. Tonelli’s project responds to frequent lamentations in the clini- cal literature about the failures of ‘knowledge translation’ (or whatever buzz word is favoured in your part of the world) [2]. Moreover, it does so by asking clinicians – in this case, an inter- national working group of intensivists – what seems to make a difference in their own decision making. This simple move has many benefits: it starts an investigation into the research–practice gap from the perspective of practice; it allows room for the pos- sibility that resistance to knowledge translation might be justified and even laudable in some cases; and it is explicitly provisional and so encourages the addition of future insights from other general and specialist physicians (and beyond). In what follows, I will clarify and provide some context for the project, identify what I see as two principal limitations of the paper, and conclude by sketching out how some of the ideas might be extended or further developed. I believe that Tonelli’s paper is a provocative starting point for discussions about the justifiability of resistance to some efforts to close the research–practice gap. Clarifications and context The results of research are compelling when they change the practice patterns of clinicians [1]. By contrast, clinical research can be assessed for its methodological quality using the widely used GRADE (Grading of Recommendations Assessment, Devel- opment and Evaluation) system [based on the hierarchy of evi- dence originally developed by proponents of evidence-based medicine (EBM)]; and guidelines can indicate the strength of a recommendation made on the basis of research depending on the uncertainties arising from a set of research results [3]. It is gener- ally acknowledged that quality and strength of guideline recom- mendations can come apart; high-quality research may lead to weak recommendations, for instance. Although Tonelli stresses the differences between constructing guidelines for general use (which is not the focus of his paper) and deciding which treatments to offer to individual patients (which is), there are some obvious similarities in these undertakings. According to the GRADE system, in the construction of guidelines, there are three types of uncertainties that can weaken a recommendation: (1) uncertainties about the balance of desirable and undesirable effects; (2) uncer- tainties about values and preferences; and (3) uncertainties about the efficient use of resources. The second and third of these factors bear strong resemblance to the second and third categories of features determining compellingness: individualized care and patient values, and stewardship of health care resources. On this reading, then, Tonelli appears to be constructing a more detailed, individualized version of the criteria that determine the strength of recommendations at the guideline level. This explains some significant similarities between the two, and also some of the differences – particularly those that arise in the first category related to epistemic factors. For instance, assessments of objectiv- ity (an epistemic factor) may lead an industry-averse clinician to dismiss a study because of concerns about conflict of interest and biased results, where producers of guidelines would most likely incorporate the study into the analysis as long as the quality of the methodology was high. Situating the project this way helps to make sense of the relationship between compellingness and some of the other terminology that arises in the paper, for instance references to valuable, relevant or trustworthy evidence. But it also raises some questions about the evidence provided in support of the arguments. Journal of Evaluation in Clinical Practice ISSN 1365-2753 © 2013 John Wiley & Sons Ltd, Journal of Evaluation in Clinical Practice 19 (2013) 559–561 559