The Day the Universe Changed - James Burke [103]
There were good political reasons for this use of the new mathematics. After the revolution, attempts at social reform on a national scale were thwarted by the simple fact that no one knew how big the population was. Planning was difficult, if not impossible. Counting every single person was out of the question, both financially and in terms of organisation. Then in 1795 the foremost French physicist, Pierre-Simon Laplace, gave a series of lectures at the Ecole Normale in Paris. His last lecture was about the calculus of probability, which, he said, he had developed through his interest in games of chance. Its use in human affairs would help eliminate ignorance of the causes of error in statistical analysis, since it was possible to reason from frequency of event to probable cause. The more frequently things happened the more could be said about their constancy and regularity of repetition.
Over the next few years Laplace went on to suggest specific uses for his calculus. He showed how it could be used to guide and improve observational methods, to evaluate the reliability of experimental results, to discover underlying natural regularities or laws hidden by irregular accidental disturbances or by large observational errors, and to suggest causes. He worked out an equation that would derive the most accurate estimated total population from an extremely small sample, and in doing so invented the concept of a statistically meaningful percentage.
The idea of using numbers in this way to improve diagnostic or therapeutic efficacy rapidly spread to the hospitals, where the multitude of patients was a prime source of large amounts of data. The earliest attempt at analysis was by the young Philippe Pinel, a friend of Benjamin Franklin’s. In 1792 he had been given charge of the Bicêtre, the hospice in Paris for the aged and infirm. It was the biggest asylum in Europe, with over 8000 patients, most of whom were considered to be beyond aid.
Pinel’s view was that slow progress was being made in medicine because inexact and untested methods were being applied. He advocated repeated observation of the sick, regular recordings of findings and comparison of data over time. This, he claimed, was the only way to arrive at the correct forms of therapy for a large number of patients. While his methods were simple, producing little more than a proportional statement of success or failure, Pinel brought public attention to the problem. His decision to remove the shackles from his patients made him a household name among his fellow-professionals.
Pinel unshackles the insane at the Salpêtrierè. A grateful patient kisses his hand. A restraining leather strap is being removed from the patient in the centre.
In the early 1820s Pinel’s methods were adopted and extended by the second head of the Paris medical school, Pierre Louis. Over a period of seven years Louis conducted no private practice at all, spending up to five hours a day in the hospital wards, gathering data on patients and then, after they died, correlating the course of their symptoms with post-mortem evidence. The surgeons had already begun doing this, but Louis’use of statistical analysis enabled him to show that his predecessors’claims of therapeutic success had been based on inexact and inadequate data. Treatment and diagnosis could now be more accurate.
Meanwhile, other advances were improving the collection of symptomatic data. The new concern with localised sites of disease aroused interest in the use of a technique originally developed by a Viennese doctor, Joseph Leopold Auenbrugger. In 1761 he had shown that tapping the chest produced sounds by which the position of the heart and the condition of the lungs could be identified. The technique was popularised by Jean-Nicolas Corvisart, Napoleon’s doctor, who was a specialist in heart conditions and founder, in 1808, of the Paris School of Morbid Medicine.
In 1816 another doctor, Théophile-René-Hyacinthe Laennec,