Author: Daniel E. Henao* and Fabian A. Jaimes**
Institution: *Grupo Reproducción, School of Medicine, University of Antioquia. Colombia **Grupo Académico de Epidemiología clínica, School of Medicine, University of Antioquia. Colombia
Correspondence: dhenao@medicina.udea.edu.co
To the Editor: Because physicians deal daily with human suffering, they should be aware about taking the best clinical decisions to mitigate their misery. This premise generates a crucial question: on what should clinicians’ decisions be based to achieve this goal? Lately, as discussed by Rosenberg et al. (1995), there is strong academic consensus that scientific evidence should be the platform that supports medical decisions; and it is precisely in this scenario where Evidence Based Medicine (EBM) is considered a new paradigm for medical practice.
EBM endeavors for taking clinical decisions combining scientific evidence from well designed clinical studies, the individual clinician experience and the patient’s preferences (Sacket et al. 1996). Commonly, clinicians used to take decisions based on their personal experience and guided by indirect evidences provided by relevant biological disciplines, such as physiology, biochemistry and immunology. In contrast, EBM proposes to use direct evidence to solve medical concerns that are out of reach of the traditional biomedical sciences such as: the time to follow-up a patient after an antibiotic treatment and the clinical predictors of mortality for patients within the emergency room. In this regard, the use of direct evidence to take clinical decisions has boosted the development of clinical studies, moving one step forward the genesis of a new "medical science" known as clinical epidemiology (Henao et al. 2009). This is a discipline that uses the scientific method to solve relevant medical dilemmas, becoming the main source of evidence that nourishes the EBM.
EBM represents an outstanding opportunity to assist physicians to achieve their goals; however a tremendous effort needs to be made to recognize that statistics and mathematics are not the only source of its evidence. The principal aim of this letter is to illustrate different potential sources of knowledge to enriched EBM.
FROM THE CLASSICAL MECHANICS PROPOSALS TO THE ILL PATIENT WITHIN THE EMERGENCY ROOM
Philosophiæ Naturalis Principia Mathematica is the title of the most exceptional work of Isaac Newton. In this publication Newton cultivated the seed of what it is considered to be the main objectives of modern science: to explain the causes of observed phenomenon through universal laws and to represent these laws in mathematical terms (Hubert 1943). The scientific basis of clinical epidemiology, and by consequence of EBM, has not been indifferent to this heritage. Sacket et al. (1996) clearly stated that one of the main advantages of EBM’s paradigm when taking medical decisions is the introduction of mathematical terms to predict an important clinical outcome. However, even these probabilistic terms are extremely important; there are other factors that should be considered when trying to diminish suffering and pain of patients; as represented in the following case.
In the clinical round of a ward, a physician presented the case of a middle-aged man who was brought to the emergency room with a stroke (cerebral bleed or infarction). Immediately, one of the medical students asked what were the chances the patient suffering from neurological disabilities. A senior neurologist answered that in all his years as a practitioner, he had seen several patients with a stroke and the probabilities of permanent neurological deficit were not that significant as long as the patient could count on family support and suitable medical follow-ups for his recovery. Nonetheless, the student was not completely convinced with this answer and surveyed the scientific literature for an alternative response. There, he found evidence demonstrating that the proportion of patients suffering permanent neurological disabilities was 15-30% (Thom et al. 2006).
What is the difference between these two answers? The quantitative terms offered by the clinical study are more precise than the answer of the neurologist. However, the answer of the neurologist was more accurate in terms of the reality for the patient and of his chances of overcoming possible sequels. Both answers are valid and complementary, and that is why answering this question just with mathematical terms would not be sufficient. In this regard, Bichat states (Canguilhem 1971) "…irregularity and instability are inherent of living systems; therefore trying to reduce them in mathematical terms would be distorting them". In this regard Andrews et al. (2003) mention that human beings, as complex living systems, are determined by their biological and mental backgrounds, their cultural context, and the dynamic relationship that exists between these components. Since there are multiple variables that interact with each component (Gebicke-Haerter 2008), those responses could not be reduced simply to probabilistic terms. Hence, the intention of clinical epidemiology to define the prognosis of a patient merely by mathematical terms has an innate epistemological limitation because the object under its study, an individual diseased person, possesses an inherent uncertainty.
The amazing successes of modern physics with respect to our knowledge of the universe, the great "enterprise" of understanding the behavior of celestial bodies and the composition of atoms, and the resulting technological progresses, should encourage us to elucidate, in an unfailing scientific manner, the dynamics that drive a diseased person to attempt to diminish his suffering and chances of dying. However, we should always be aware of the fact that the legacy that Newton left us when representing the movement of the celestial bodies thorough an equation is not enough when facing a human being in pain in an emergency room. As in the case of this patient, although his chances of becoming neurologic impaired could be of 15-30%, his fears, hopes, dreams and expectations cannot be mathematically represented.
TO BE SICK OR NOT TO BE SICK: That’s the dilemma
The first obstacle the researcher should overcome when designing clinical research is: who will be considered "patient" and who "no patient", or even better who will be considered ill and who healthy.
When designing his thesis research project, a first-year PhD student was really interested in carrying out a double-blind randomized clinical trial to diminish hospital admissions of patients diagnosed with heart failure and undergoing pharmacological and physical treatment compared to those who only received pharmacological treatment. His first step in designing his research was to define the population to be studied. To do this he guided himself with the definition of heart failure proposed by American College of Cardiology/American Heart Association (Hunt et al. 2005) which states that "(…) heart failure is a syndrome associated with cardiac dilation and impaired cardiac contractility". However, this definition would not solve his main question: how would he define who is sick and who is not? Although the main clinical representation of an impaired cardiac function is a diminished ejection fraction measured by an echocardiography, is an ejection fraction less than 50% enough to define that a patient is ill?
In trying to clarify this important question we should briefly discuss the difference between the normal and pathological states. Comte (Hubert 1943) once assumed that "…the pathologic state does not differ in its essence from the normal state; it only constitutes a mere prolongation of its normal equivalent". The corollary of Comte’s statement could be that disease and health can be homogenized in continuous terms. However, although mathematical representations of clinical outcomes are extremely valuable, they are limited because there are constructs based on categories within a human being and cannot be considered continuous. One can consider the following example: two very good friends found themselves starving and went out for dinner. Once in their home they faced a paradoxical situation: in total they ate 10 loaves of bread, which means that on average each one would have eaten 5 loaves. However, one of them was still starving even though he had eaten 6 loaves and the other was full even though he had eaten 4 loaves. What can we learn from this example? Mean, medians, and standard deviations alone cannot define the presence of an emotion In the same manner these numerical associations cannot define the diagnosis of a disease because a pathological state is in a different category than the normal state and, more importantly, this "abnormal" condition may be extremely variable among and between individuals. Illness, diseases, health and complaints cannot be homogenized on a quantitative scale. In other words, statistical associations that arise from clinical studies are necessary to make clinical decisions but they are not sufficient because there are hundreds of variables that the clinician should consider to diminish suffering and pain. In this respect, the importance of heart failure cannot be represented merely by the percentage of ejection fraction. For a patient to suffer from heart failure there should be a systemic reorganization process in which his heart could not respond coherently to metabolic demands; and from this qualitative situation emerges the disease, rather than from a quantitative distortion.
In conclusion, we should accept that the progresses achieved by the medical sciences in the last 100 years are unimaginably important, but we should remember that the practice of medicine contains a lot of art. Furthermore, sufferings and death of human beings are, sometimes, out of reach of any kind of evidence.
REFERENCES
Andrews G, Halford GS, Bunch KM, Bowden D, Jones T. (2003) Theory of mind and relational complexity. Child Dev 74(5): 1476-99
Canguilhem G. (1971) The normal and pathologic states. First edition. 21TH Century
Gebicke-Haerter PJ. (2008) Systems biology in molecular psychiatry. Pharmacopsychiatry 41 Suppl 1: S19-27.
Henao DE, Jaimes FA. (2009) Medicina Basada en la Evidencia: Una aproximación epistemológica. Biomédica 29:33-42.
Hubert R. (1943) Philosophical breviaries "Comte". First edition. South America editorial
Hunt SA, Abraham WT, Chin MH, et al. (2005) ACC/AHA 2005 guideline update for the diagnosis and management of chronic heart failure in the adult: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation 112: e154-e235.
Rosenberg W, Donald A. (1995) Evidence based medicine: an approach to clinical problem-solving. BMJ 310(6987): 1122-6.
Sacket DL, Rosenberg WM, Gray JA, Haynes RB, Richardson WS. (1996) Evidence based medicine: what it is and what it isn’t . BMJ 312(7023): 71-2
Thom T, Haase N, Rosamond W, Howard VJ, Rumsfeld J, Manolio T, et al. (2006) Heart Disease and Stroke Statistics- 2006 Update. A Report from the American Heart Association Statistics Committee and Stroke Statistics Subcommittee. Circulation 113: e85-e151