Espistemic Humbleness versus Epistemic Arrogance… Statistician versus clinician.
Resorting to a thermodynamic metaphor, we can imagine the clinical evolution of the single patient as the trajectory run by the single molecule of gas. The statistical approach allows to make assertions about some macroscopic characteristics of the mass of a gas (e.g. average kinetic energy, temperature, pressure, etc.). These mean values may be modified administering, for example, some quantity of heat to the entire mass of the gas. It is the very sense of a statistical experiment, as an RCT (Randomized Clinical Trial): we administer in "watering way" (all-round distribution, “rain-distribution”), a specific treatment to a mass of individuals (supposed to be independent each other) and, then, we are going to evaluate the modification of some parameters, calculating the mean, standard deviations and comparing the means between the groups. Sticking to the above metaphor, the clinician is, instead, interested to modify the trajectory of a single molecule and confine it in a certain space region.
Thus, it would seems that an apparent irreducibleness of the clinical situation into the statistical studies. The clinician and the statistics, yet interested to the same problem, would observe the patients from two different points of view, using different temporal and spatial scales, aiming different objectives.
Ultimately, they attempt different levels of descriptions: the statistician would use a “macroscopic” description, the clinician would use a “microscopic” description.
In truth, the interconnections and interactions between the statistician and the clinician are very tight: the macroscopical informations deriving from the statistical analysis influence the microscopical behavior (adressed to the single patient) of the clinician, who, in its turn, receiving feedbacks from following the patients, formulate new working-hypothesis to be statistically analyzed, and so on.
The following hypothetical, yet likely, story may be useful to comprehend the very nature of the problem, single versus collective, macroscopic versus microscopic:
"Mister Rossi goes out in the morning for meeting the doctor owing to an urinary problem. Before leaving the building , the janitor hold him for reminding the payment of some residents' bills. Once come out from the building, Mister Rossi is run over by the car of Mister Mori. Mister Mori went fast with his car because, due to the traffic encountered, he was risking to loose an important appointment for his career advancement. Mister Rossi got a cranial trauma and goes into coma before ambulance come. Once he arrived in hospital doctor cannot do any tests before 1 hour because too busy. Mister Rossi dies”
We could conclude that the death of Mister Rossi were caused by the anoxia of brain cells as consequence of the haemorrhage caused by the trauma. But the cerebral anoxia may be advocated for all the deaths: every cerebral death happens when the oxygen supply ceases to the brain!
Or can we attribute the death of Mister Rossi to the trauma? Or can we attribute it to the hurry of Mister Mori? Or to the urinary disturbances of Mister Rossi? Or to the janitor who had been holding Mister Rossi for some minutes? Or to the retard of the Ambulance? Etc.
As you can note, many levels of description of the causes are possible and plausible. The plausibility of these etiological explanations derives from the respect of the so called “counterfactual reasoning”: Mister Rossi would have not died if he did not suffer of urinary disturbances, or if did not get an appointment with the doctor, or if Mister Mori did not have career problem, etc.
Each of these “if not” configure a possible world, which if had ever occurred, it would have impeded, perhaps, the death of Mister Rossi due to that incident, at that time, in that place!
The problem is that these “potential worlds” are inaccessible to our “experience”: there is no way to demonstrate the validity of a counterfactual reasoning, being realized solely a future among all the possible futures: that future ending with the death of Mister Rossi run over by the car of Mister Mori.
It is reasonable to suppose there is no substantial difference between our fictitious example and the most commonly encountered medical situations, as for instance the occurrence of bacterial infections, or the occurrence of a cancer, etc.
Counterfactual reasoning is practically ubiquitous in medicine:
– you would have never caught that infection if your immune system had had a former contact with the microorganism… if you had no met the infected person, and of course if you had not gone to that place
– It’s frequent to hear, about a person healed after a surgical operation, that he would have certainly died if he had not been operated in time
– it’s frequent tell to a patient: if you had no committed this imprudence, you would have not been in this condition
– ….if you had not shaped the forth canal in this first upper molar, the endodontic treatment would have failed
-…..if you had not got to the apex, the endodontic treatment would have failed
– and so forth
From these statements you can infer the great hazard hiding behind the counterfactual reasoning: every event could be explained by means of a series of “if not”, whose validity is not easily demonstrable.
I wish to conclude citing this pass of interview to Primo Levi, Italian chemist and writer, prisoner of Nazi Camp during the second world war. During the interview, the journalist asked to Primo Levi the following questions :”What would have you, today, if you had not been prisoner of Nazi Camp?…..Which factors do you bestow your survival to?
The Primo Levi’s answer was:
“Rigorously speaking, I do not know and I may not know what I would have been today if I had not been in Lager (Nazi camp); no man knows his own future, so I cannot describe a never-been future. Attempting some prevision, yet roughly, on the behaviour of a population has a certain significance , but it is impossible to forecast the behaviour of an individual, even in a daily-scale. At the same way, a physicist is able to foresee with extreme exactness the time of decay of a gram of radium, but he’s not able to predict when a single atom of that radio shall disintegrate. When a man reach a fork, and he does not enter the left road, it is obvious that he turns into the right road. Hardly ever, our choices are among just two alternatives as in the case of the fork. Then, to each choice another one follows, all multiple, and so forth ad libitum; ultimately, our future is strongly influenced by external factors, of whom we are not aware. Because of all these reasons we may not know our future nor the future of our neighbour; for the same reasons no man can say what would have been his own past if………….”