Association Does Not Equal Causation—Does Thinking You Are Better Make You Better?

Appendicitis, traditionally framed as a mechanical obstruction of the lumen causing infection, is now considered to be biological (or microbiological) given that, in select cases, it can be successfully treated with antibiotics. Therefore, under this new framework, it is reasonable to hypothesize that mind-body connections can influence the infectious process based on emerging evidence in this field. For example, holistic medicine claims that the mind, body, spirit, and emotions of a patient need to be addressed for healing to occur,¹ proclaiming “if you think you are better, then you are better.” The well-done study by the Writing Group for the CODA Collaborative, using a secondary analysis of 425 patients in the Comparison of Outcomes of Antibiotic Drugs and Appendectomy (CODA) trial who were not aware of their treatment assignment at the time of the survey, tested whether belief in the success of a given treatment choice in appendicitis can influence outcome.² They concluded that there exists an association between patients’ beliefs and outcomes for appendicitis treatment (i.e., via antibiotic administration). Yet whether such associations can advance our understanding of the many factors involved remains unclear. As the authors aptly recognize and as we are all aware, a purely descriptive study such as this cannot clarify whether an observed association implies causation. While investigators in the field of mind-body biology have generated compelling evidence for this type of mind-body connection,³ linking their work to support the observations made in the present study is problematic as it suffers from syllogistic reasoning, the “between-group comparison fallacy,”⁴ and confuses association with causation. Although the authors of this laudable effort explicitly acknowledge this bias, readers must be aware that rigorous molecular-based studies are needed to shift such descriptive-level observations from remaining probabilistic so they become more deterministic. Many molecular-level interactions must be accounted for in order to identify the determinants that are causally responsible for the observed clinical effect. These include the attitudes of the physicians, nurses, family members, etc, the pheromones they emit and share with one another, how these transferred signals are received and responded to, and how this giant “interactome” mechanistically influences outcome.⁵ Absent this information, we are left with associations that cannot be determined to be either causative nor actionable. While the authors should be congratulated for generating information that suggests there “might be something here,” the hard work needed to support mechanistic plausibility to this claim, as they acknowledge, remains to be produced. Unless we insist on complementary biology-level measurements and hard mechanistic-based evidence to support our clinical observations, it seems we are no better than homeopathic claims that declare “if you think you are better, then you are better.”