The Paradoxes

Infantile strabismus signals a neurodevelopmental aberration in the developing visual system. For the better part of a century, a nature (Worthian) versus nurture (Chavessian) debate has framed our concept of causation in infantile strabismus. Infantile nystagmus in children is dichotomized into those disorders known to be associated with a sensory visual disorder and those that occur in the absence of any known deficit. As new genomic, proteomic, and neurobiological mechanisms of disease are discovered, a simple mechanistic explanation for each of these conditions remains conspicuous by its absence. We are confronted with the paradox that, as medical science becomes more precise, causality seems to recede.

To establish causality, one first must understand scientific law. But to formulate scientific law, one first must understand causality. Such is the epistemological paradox of scientific thinking. In pediatric ophthalmology, assigning causation has been impeded by several problems. First, it is easy is to confuse links and correlation with cause and effect. Second, we have an inherent tendency to conflate form and function. When a convergent deviation of the eyes is the sine qua non of infantile esotropia, it may be hard to imagine that this condition may not be caused by excessive convergence. Third, terminology can potentiate misinterpretation of causality. Because of commonly used terminology (pursuit eye movements, fixational eye movements), we tend to forget that pursuit and fixation are the effect of eye movements rather than the cause. Finally, we are trained to think linearly about problems that fundamentally are circular. In so doing, we lose sight of our basic physiology as a complex feedback loop. In such a system, cause and effect can change purely as a function of when you jump on the loop. The problem is that the effect becomes the cause; for example, intermittent exotropia may lead to tight lateral rectus muscles, which then become causal.

This discussion addresses the complex problem of interpreting causality in the evolving era of systems medicine. According to Federoff and Gostin, “Systems biology is based on a network of interacting components—such as the coordination of internal systems (nervous, endocrine, respiratory, etc) with gene and gene product expression and behavioral and environmental factors—and how these components contribute to the course and health of disease.” Systems medicine “incorporates attributes uniquely human, such as human genetics, environment, and behavior.

The Precedent

In a related field, the law assigns causation in a way that is different but instructive and relevant. In the courtroom, it becomes imperative to assign a cause or a value to a purported injury. A diabetic steps on a nail, a foot ulcer develops, and the foot has to be amputated. What was the cause? The rusty nail? The diabetes? The irresponsible workman who left the rusty nail in his path? It is true that he probably would not have gotten the infection without the diabetes, but if it were not for the nail, he would still have his leg. In assigning responsibility or liability, the legal system distinguishes proximate (direct or immediate) cause from other but for causes (necessary conditions in that particular situation). Only the proximate cause determines whether the injured party deserves compensation. The nail provided the mechanism of injury, so the workman’s negligence was the proximate cause of the loss of the leg. In the law, you take the victim as you find him, so the contribution of the diabetes to the injury does not diminish the extent of the workman’s liability.

The Precedent

In a related field, the law assigns causation in a way that is different but instructive and relevant. In the courtroom, it becomes imperative to assign a cause or a value to a purported injury. A diabetic steps on a nail, a foot ulcer develops, and the foot has to be amputated. What was the cause? The rusty nail? The diabetes? The irresponsible workman who left the rusty nail in his path? It is true that he probably would not have gotten the infection without the diabetes, but if it were not for the nail, he would still have his leg. In assigning responsibility or liability, the legal system distinguishes proximate (direct or immediate) cause from other but for causes (necessary conditions in that particular situation). Only the proximate cause determines whether the injured party deserves compensation. The nail provided the mechanism of injury, so the workman’s negligence was the proximate cause of the loss of the leg. In the law, you take the victim as you find him, so the contribution of the diabetes to the injury does not diminish the extent of the workman’s liability.

The Problem

In medicine, the distinction between proximate cause and mechanism becomes almost tautological. Is an inherent instability of the developing pursuit system, which has been postulated to cause infantile nystagmus, the proximate cause or the mechanism? Perhaps an undamped smooth pursuit system is the mechanism and the failure to calibrate it is the proximate cause, but this is a fine hair to split. The infinitude of cellular and molecular processes in medical disease makes it convenient, expeditious, and indeed necessary to assign the potentially modifiable risk factor as the but for cause in any given patient.

Infantile nystagmus may be caused by genetic mutations or congenital visual loss, but for which it would not have been expressed in that particular child. In a court of law, however, if the pediatric ophthalmologist fails to remove congenital cataracts promptly and the parents sue because infantile nystagmus develops in the baby, the physician who failed to remove them could be held liable if his failure to act expeditiously were deemed the proximate cause of the nystagmus. If he tried to argue that the proximate cause was an undamped pursuit system, he probably would lose, which shows how imprecise and utilitarian our concept of proximate cause really is.

The Patient

For the patient, medical diagnosis often is synonymous with cause. The patient with a cough may consider bronchitis to be the cause, when in fact it is the medical diagnosis, and the actual cause is bacterial infection. I recently overheard a patient announce, “I just found out that I have Wilson’s disease. I’m so happy, I finally have a diagnosis.” For the patient, receiving a diagnosis may impart comfort or security by explaining the origin of physical symptoms, creating order out of chaos.

The Physician

As physicians, the way we assign causation imparts our longstanding cognitive biases to the complexity of events that occur within dynamic systems. Our human tendency to construct a unidirectional view of causality can lock us in to a closed system of thought. Many are familiar with the perceptual phenomenon of the Necker cube, wherein which a frame drawing of a cube in oblique perspective appears to flip back and forth between 2 3-dimensional cubes. Quantum leaps in scientific understanding can come from a so-called flipping of the Necker cube in a way that transposes cause and effect. The Copernican system of astronomy and the Darwinian system of evolution are but 2 examples of causal paradigm shifts that have generated fundamental inversions of our subjective reality by necessitating a reversal in our view of causality.

In medicine, our search for causation is focused primarily on identifying but for (or necessary) causes, which include any modifiable risk factors that can be prevented, circumvented, or treated. Infectious diseases such as tuberculosis, syphilis, and (even) duodenal ulcer conform mercifully to this paradigm, as does the cause-and-effect association of lung cancer with smoking. When we hold the magnifying glass over many diseases, what comes into focus are multiple predisposing risk factors (genetic and environmental) of relative weight. Consider age-related macular degeneration, where time (aging) acts in synergy with multiple physical and environmental causes (genetics, cholesterol, light exposure, inflammation, ischemia, smoking). Even epigenetic factors (heritable changes that alter the expression of a gene without changing the DNA sequence) are now in the mix. We have discovered that common hereditary diseases can have multiple independent genetic causes, as in Leber congenital amaurosis. For the physician, common conditions may require further investigation (e.g., genetic investigation for mutations in tumor suppressor genes).

When dealing with but for causes, the tail of the comet is long. In medicine, remote or permissive causes become important modifiable risk factors. According to the Barker hypothesis, causation exists along the 3-dimensional axis of physical events, environmental events, and time, so that poor nutrition early in utero or in childhood is a risk factor for developing heart disease, high cholesterol, stroke, and many other diseases in adulthood. This is no easy journey, because our causal inquiry into disease must weigh the effects of multiple diseases and predisposing factors (teratogens, sun exposure, socioeconomic conditions, environmental toxins) along with negative experiential factors (failure to exercise, to maintain a balanced diet, to obtain enough sleep, to take vacations). As the footprints of causality are traced further and further back in time, any definition of cause (epidemiologic, mechanical, physiologic, environmental, molecular, genetic, epigenetic, etc.) necessarily becomes a function of one’s reference point.

Although we lean heavily on the word etiology (the study of causes and origins of disease), it is also pathogenesis (the cascade of events leading to the development of a disease) that we are after. Pathogenesis is closely related to mechanism. Consider the neurovascular phenomenon of migraine, which represents a system malfunction that occurs via a cascade of events. The centerpiece is a susceptibility to recurrent headaches that are initiated by various environmental triggers. Affected patients may have inherited one or more genetic polymorphisms that, somewhere along the cascade of events, cause instability. It now seems that migraine aura and its associated headache can be caused by either inappropriate modulation of normal trigeminocervical pathways or normal activation of low-threshold trigeminocervical pathways. Add to this the fact that a biological system is dynamic and that a repeatedly activated system changes over time, and it becomes easy to understand how a problem that has its roots in genetic susceptibility can evolve into a chronic form that obscures its underlying nature. Disease causes disease, which is why it is generally good to stay healthy.

In their critical analysis of perinatal brain damage, Dammann and Leviton addressed 9 reasoning errors in medical causation that impact our understanding of disease:

• 1
  

  Too narrow a view of what is a cause (causal production vs facilitation)

  
  
• 2
  

  Extrapolation from possibility to fact (potential vs factual causation)

  
  
• 3
  

  If X, then invariably Y (determinism vs probabalism)

  
  
• 4
  

  Co-occurrence in individuals versus association in populations

  
  
• 5
  

  One cause is all that is needed (single cause attribution vs multicausal constellations)

  
  
• 6
  

  Drawing causal inferences from very small numbers of observations (the tendency to generalize)

  
  
• 7
  

  Unstated causal inferences

  
  
• 8
  

  Ignoring heterogeneity

  
  
• 9
  

  Failing to consider alternative explanations for what is observed.

This checklist provides a useful template for the physician to navigate and critically question the often biased conclusions that arise from evidence-based studies in the systems era of medicine.