The King Is Mad; Long Live Statistics
The collection and manipulation of data, so central to the study of human heredity, knows no unique point of origin. The assembly of a committee of the House of Commons in January 1789 to inquire into George III’s apparent madness, potentially a crisis situation, brought to the fore some vital questions about proper recordkeeping. A highly politicized disagreement about the king’s medical prognosis had roused a furor among the royal physicians extending to Parliament and the nation. How could they anticipate whether he would recover in time to preclude any need for a regency? What confidence was to be placed in the Reverend Dr. Francis Willis, who had been brought in from the provinces to care for the king? The committee questioned Dr. Richard Warren, one of the king’s physicians:
Whether if Nine Persons out of Ten, placed under the Care of a Person who had made this Branch of Medicine his particular Study, had recovered, if they were placed under his Care within Three Months after they had begun to be afflicted with the Disorder, Doctor Warren would not deem such Person, either very skilful or very successful?
Warren answered yes, if so many really did recover.
Whether, in order to induce Doctor Warren to believe, that, for Twenty-seven years, Nine persons out of Ten had been cured, he would not require some other Evidence than the Assertions of the Man pretending to have performed such Cures?
I certainly should.1
Willis, having made very specific claims, could not back them up with written records. Dr. Warren’s sharp skepticism points to an emerging ethic of data and accountability as a basis for public knowledge. The administration of madness participated fully in an explosion of statistical activity across a range of scientific, bureaucratic, and professional projects that already was beginning to take shape. Public numbers were not only for legislators but also for common people, who were urged to calculate in order to judge their rulers as well as to better their lives. The debates sparked by Thomas Malthus’s Essay on Population in 1798 were unprecedentedly lively, if far from new. Medicine, too, experienced sharp controversies, most famously regarding the advantages of inoculation campaigns against smallpox. Many physicians, however, were not eager to be judged by their numbers. If patients and diseases were unique, undiscerning data could not reliably guide a treatment decision. A public campaign for “numerical method” took off finally in the 1820s, launching an era of statistical hope coupled with doubts about the old remedies.2 The treatment of the insane was more profoundly reshaped by numbers than was ordinary medicine. Asylum medicine, as an area of public health, had only a weak ethic of individualism. In an age when physicians were paid directly by patients they visited in their homes, hospitals were more like poorhouses. Much confinement of the mad was for the sake of public safety, and many patients, lacking the means to pay the costs of their treatment, were designated paupers. As public or charitable institutions, asylums were particularly vulnerable to accountability standards, which meant providing numbers for patient admissions and outcomes as well as for revenues and expenditures.3
Asylum statistics originated as a form of bookkeeping. The prototype of the patient reckoning was a balance sheet. John Strype’s 1720 edition of Stow’s classic survey of London includes tables for Bethlem Hospital (Bedlam) beginning in 1704. The “Disturbed Men and Women then brought in” for the year numbered 64, while 50 were “Cured of their Lunacy and Discharged,” and 20 more were “Buried,” leaving 130 patients under care. The table for 1705 shows 72 admissions, 34 cures, 29 deaths, and 137 patients remaining. (This arithmetic leaves 2 patients unaccounted for.) The next year, “as it was Published,” is “1705 to 1706,” with 72 patients admitted, 52 cured, and 13 buried. Strype gave another table extending from Easter 1706 to Easter 1707, followed by three years with no information, a table for 1711, five more unrecorded years, and tables for 1717 and 1718.4
If the accounting seems fragmentary, the therapeutic evidence was still more so. The reports from 1680 to 1705 claimed cure rates between 57% and 82%. Yet the patients discharged as cured were not distinguished in the records from those sent away still mad. Some, certainly, did not recover, for Bedlam had an announced policy to limit the residence of patients to about a year. (The evidence of the numbers, with annual admissions about half as numerous as patients remaining under care, implies an average stay of about two years.) Strype did not explain the calculation of cure rates, merely proclaiming: “So that by God’s Blessing for Twenty Years past, ending 1703, there have been above two Patients in three cured, as the Physician hath told me.” Yet Bethlem was in some manner held responsible for patient outcomes, and the higher powers were not so lax as to leave the numbers to happenstance. Hence the need for techniques of deception. A Bethlem historian remarks that they maintained a low death rate by proactively discharging weak or debilitated patients.5
Although many asylum officials continued to treat their records as proprietary into the nineteenth century, others were working to expand public access to data on the mad. Dr. William Black, who was trained up in medical arithmetic by the campaign for smallpox inoculation, had complained in 1781 that at Bethlem, the “relieved, cured and discharged, are jumbled into one list” so that none but its “eminent physician,” Dr. John Monro, could resolve what proportion is cured. Was this a backhanded compliment? In 1788, Black described the Bethlem medical books as an “untrodden wilderness.” The eminent Dr. Monro, however, introduced him to his estimable son, Thomas, who in turn led Black to yet another praiseworthy individual, the resident apothecary, John Gozna, “whose learning and curiosity induced him to keep a private register of all the patients, upon which, as incontrovertible data, I have founded and collected all the following tables and propositions.”6
At this point, the king’s madness turned the barren wilderness of Bethlem records into terrain of extraordinary value. Black, quite unexpectedly, found himself holding the key to unlock a great medical-political mystery. He set to work in great haste on a new edition of his just-published work on human mortality, withdrawing the 1788 version from circulation. In the crucial year of George III’s madness, he offered a solution to the urgent problem of prognostication. In a dedication of his book to the younger George, Prince of Wales, who was more than willing to assume the regency that Black’s numbers were likely to advance, Black declared: “I trust it will not be arrogant in me to say, there will be found considerable original, useful, and authentick information.” Such was the power of the medical numbers he prepared from Gozna’s record book. “I may with safety assert, that mine are the only numerical and certain data that ever have been published in any age or country, by which to calculate the probabilities of recovery, of death, and of relapse in every species and stage of insanity, and in every age.” The tools of arithmetic, so valuable “in politicks and commerce,” would now prove themselves in medicine. Insanity, long written off as “the most difficult and conjectural” topic in all of medicine, would at last yield to “medical arithmetick.”7
These were heady times for Black. “And it is not a little flattering, that the interrogatories, in thi...