We are pleased to welcome an insight on narratives, values, and medicine by Dien Ho, PhD, Associate Professor of Philosophy and Healthcare Ethics, Director of the Centre for Health Humanities at Massachusetts College of Pharmacy and Health Sciences (MCPHS) in Boston. His A Philosopher Goes to the Doctor: A Critical Look at Philosophical Assumptions in Medicine has been recently published, and we hosted his interview in July 2019.
The indispensable role of narratives in clinical medicine has been well-established. Care providers who have been trained in narrative medicine exhibit a higher degree of compassion and empathy in their interactions with patients. Nonetheless, narratives enter medicine in areas that are often less noticed. In this short piece, I wish to highlight two places where narratives can affect both research and clinical medicine in surprising ways. The upshot is that training in narrative medicine not only improves bedside outcomes, it also provides an insight into the structure of medicine and, more importantly, how we understand ourselves.
Narratives as the scientific kingmakers
The discussion of science vs. pseudo-science is of more than philosophical interests. Pseudo-scientific practices are regularly shunned by mainstream medicine; reputable journals refrain from publishing studies that hint at spooky interactions, grant agencies are unlikely to fund research in exotic therapeutic modalities, and insurance companies and governmental agencies are hesitant to cover treatments with seemingly implausible causal pathways. Their reluctance comes not from the lack of evidence that these exotic treatments are therapeutically beneficial; after all, plenty of treatments in conventional medicine that are known to be of marginal benefits (if not plainly useless) continue to be supported [1]. The reason why exotic treatments are excluded has to do with the fact that they are incompatible with our empirical view of physiology and pathology nested within a larger mechanistic scientific view of the world.
Take the therapeutic benefits of remote intercessory prayers. If one were to sketch the supposed therapeutic pathway of these prayers, it would be something like the following: a person, who is hundreds of miles away from a patient, prays to a non-physical divine entity. The entity then intervenes in a non-physical manner that causes an improvement in the physical outcomes of a patient. Prayers as medicine seems outlandish not because we have no evidence for its effectiveness for even if there is evidence from clinical trials, it will likely be dismissed as data noise. The crux of the skepticism lies instead in the absence of a plausible (causal) story that can link prayers to clinical outcomes. Modern medicine operates within a scientific world view that roughly connects happenings via physical causal connections (quantum phenomena, notwithstanding). If there is no obvious way to draw a causal connection from treatments to outcomes, whatever clinical evidence one can obtain will likely be chalked up as coincidental.
Vertebroplasty offers an example of the power of narrative. As a standard treatment for fracture vertebras, vertebroplasty paints a commonsensical picture. With the help of a simple x-ray, surgeons can locate the crack in a patient’s vertebra—a crack that resembles those one might find in the structural supports of our homes. And, just as we might repair a crack in the foundation of building by injecting cement to restore structural support, surgeons inject organic cement into the cracks of a fractured vertebra. The clinical results are impressive; in one study, the average pain level (as identified by a 0-10 visual analog scale) dropped from 8.4 pre-procedure to 3.2 two years post-procedure [2].
Dr. Kallmes et al. [3] of Mayo Clinics’ Department of Neurointerventional Radiology noticed that occasionally surgeons inject the bone cement into the incorrect vertebra and, remarkably, patients report improvements even though the original vertebra remains untouched, fractured and all. Kallmes et al. proceeded to investigate the therapeutic benefits of vertebroplasty by running a single-blind randomized trial in which half the cohort received the verum version and the other half sham which consisted of no more than opening containers of cement that smells like nail-polish remover with no penetration by the needle. The patients from the sham arm performed just as well as the verum arm in terms of pain intensity and Roland-Morris Disability Questionnaire. Even after his study was published, vertebroplasty continues to a standard treatment for fractured back. In the United States, vertebroplasty was performed over 6,130 times in 2014 [4]. I suspect the initial embrace and continual recalcitrance are likely the mental story that we form when presented with the physical description of vertebroplasty. There is a crack, we fill in the gap, and you will feel better. It is a likely story.
Whether a potential or existing intervention is taken seriously in conventional medicine depends on the ability of its proponent to draw a plausible causal connection between the treatment and the outcome. The growing acceptance of acupuncture as a plausible treatment, for instance, is probably due to its causal plausibility. Who knows if qi is modified in the placement of needles? But, the fact that qi could exist and that needles could causally affect its flow give acupuncture the (cover) story necessary to gain respectability and plausibility. Acupuncture might or might not provide therapeutic benefits but once the rough causal story is established, we see that it could be a genuine therapy. This sketch of a causal story suffices for its inclusion into research and clinical medicine. In this sense, storytelling can make or break a therapeutic approach [5].
Narratives as the evidential kingmakers
A second and related area where narratives play a critical role in shaping medicine is at the level of evidential support. Although Evidence-Based Medicine (EBM) is typically thought of as a recent movement in clinical care, obviously no self-respecting clinician would recommend a therapy when there is no evidence whatsoever that it can benefit her patient. The novelty (and controversy) of EBM rest in its ranking of the quality of diverse types of evidence. In its crudest form, EBM ranks randomized clinical trials (and meta-analyses of them) on top of the evidential pecking order. Epistemic criticisms of EBM have been numerous and forceful; however, critics have often overlooked a fundamental internal tension in the EBM movement.
What EBM seeks is a regimentation of evidence with the hope that once we have sorted the relative quality of different kinds of evidence (e.g., RCT is better than observation trials), we will find a medicine that is more effective and less vulnerable to the subjective biases of care providers. But as any post-positivist scholar in epistemology knows, what makes a piece of data evidence for a hypothesis is not something that can be determined by the logical and semantical relationships between evidence and hypotheses. The hypothesis Smoking causes cancer enjoys mountains of evidence from clinical trials to epidemiological studies. Yet, we also have mountains of positive instances for the hypothesis that no one over the age of 150 has ever crossed the Seine. Indeed, the odds are good that there are more instances of people under 150-year old who have traveled across the Seine than smokers who have experienced a rise in cancer risks. Of course, no one in their right mind would infer that this exceptionless regularity is evidence at all for the hypothesis that anyone who crosses the Seine must be less than 150-year-old. Whether an observation qualifies as evidence for a given hypothesis, thus, depends on more than being a positive instance of the hypothesis. We must determine antecedently that a hypothesis is in fact susceptible to being confirmed by positive instances. Unlike Smoking causes cancer, Anyone who crosses the Seine must be under 150-year-old lacks the necessary nomological properties that render it capable of being confirmed regardless of how many positive instances we collect. The reason, many philosophers believe, is that the Seine hypothesis is not a proper candidate for a law of nature: there seems to be no (causal) necessity between one’s age and one’s possibility of crossing the Seine.
Here, we see again how a causal narrative can affect the working of science. Our hesitation in investigating whether age and the Seine are somehow metaphysically connected stems from the implausibility of a causal connection between the two. How could being of a certain age entails that one is metaphysically (not physically) incapable of crossing the Seine? It is not as if a person of 150-year-old would somehow magically be prevented from crossing the Seine by the laws of nature. In order for a hypothesis to be a “live” option (to borrow William James’ term) and enjoy evidential support, it must fit into our larger folk causal understanding of the world. No plausible story linking age and Seine-crossing, then no possibility of evidential support.
Both of the examples that I all too briefly discussed tell us that narratives show up in unexpected places within an objective practice of scientific medicine. And, who among us are more masterful at understanding narratives than those who are trained in literature, the arts, and the humanities? We tell stories to make sense of the world around us. We do so not because we are in need of psychological solace in an uncertain world; rather, narratives serve as, to borrow Kant’s phrase, “conditions of possibility” for understanding the world. In other words, there is no way we can engage the world without presupposing some story. The introduction of narratives into medicine is not merely to improve research and clinical care. It is instead a long overdue recognition that medicine, like life, has always been possible only against the backdrop of shared stories of how the world works. Discovering these narratives gives us the chance to shape medicine in accordance with our common values. We need not sit idly by as healthcare professionals explain and direct our care; we can contribute to this social practice by exploring and directing the tacit stories. After all, these stories are our stories.
[1] Among medical practitioners, the folk wisdom is that 1/3 of medical treatments at any given time are useless. The tricky part is identifying which third.
[2] Layton, K. F., Thielen, K. R., Koch, C. A., Luetmer, P. H., Lane, J. I., Wald, J. T., & Kallmes, D. F. (2007). Vertebroplasty, First 1000 Levels of a Single Center: Evaluation of the Outcomes and Complications. American Journal of Neuroradiology, 28(4), 683-689.
[3] Kallmes, D. F., Comstock, B. A., Heagerty, P. J., Turner, J. A., Wilson, D. J., Diamond, T. H., . . . Jarvik, J. G. (2009). A Randomized Trial of Vertebroplasty for Osteoporotic Spinal Fractures. New England Journal of Medicine, 361(6), 569-579. doi:10.1056/NEJMoa0900563.
[4] Laratta, J. L., Shillingford, J. N., Lombardi, J. M., Mueller, J. D., Reddy, H., Saifi, C., Lehman, R. A. (2017). Utilization of vertebroplasty and kyphoplasty procedures throughout the United States over a recent decade: an analysis of the Nationwide Inpatient Sample. Journal of spine surgery (Hong Kong), 3(3), 364-370. doi:10.21037/jss.2017.08.02
[5] For those familiar with the long history of Humean skepticism towards causation, one might see commonsense causation as a form of storytelling. What distinguishes constant conjunctions that are accidental from those that are linked by genuine causal connection does not rest on some metaphysical reality (i.e., there really exists some unobservable causal glue between the pool ball being struck and its trajectory into the corner pocket). Medical plausibility as storytelling then becomes a matter of demonstrating how a novel story satisfies the standards of what counts as a good (metaphysical) story. On the topic of storytelling, of course medicine is not unique in that sense. Plenty of well-respected scientific discipline depends on explicit storytelling. Take evolutionary biology. One task for those in favor of natural selection as the sole mechanism by which biological features like eyebrows arose is to offer a plausible story of how small evolutionarily steps can add up to the bushy eyebrows we have now. If such a story cannot be told (e.g., there simply wasn’t enough time to gradually evolve eyebrows), then the viability of the theory within natural selection diminishes.