The Road to Medical Statistics (review)

This is not a long book. Made up of only six essays held together by a brief but incisive introduction, it nevertheless makes a useful contribution to the burgeoning literature on medical quantification. After briefly discussing existing literature on the subject (from a British and American perspective), the editors introduce an important distinction between simple quantification (what I would call "counting") and statistics using more sophisticated mathematical techniques. The essays that follow can be seen as exemplifying the movement from the first to the second in the British medical context.

Phillip Kreager gets things started with a long and erudite essay on John Graunt's late seventeenth-century work on the London Bills of Mortality, often seen by historians as one of the first works of vital statistics. Kreager, however, sees this work as belonging to rather different traditions: Baconian natural history on the one hand, and what he calls "mercantile bookkeeping" on the other. If this was most definitely not "statistics," it was recognized, in Kreager's account, to be a "new way of reasoning of a general and potentially scientific kind" (p. 4). Perhaps the most original part of this analysis is the author's effort to place this new method within the framework of rhetorical analysis. Numbers, in other words, were not just mathematical entities but part of the ordinary language that sought to understand and convince.

Andrea Rusnock looks at the work of Thomas Nettleton to similarly argue that eighteenth-century discussions about the risks and benefits of smallpox inoculation introduced a new kind of quantified argumentation in medicine. Like Kreager, she emphasizes the nonmathematical quality of this reasoning by calling it "Merchant's Logick." This method of argumentation, she suggests, became increasingly common in discussions of inoculation during the eighteenth century.

Edward Higgs presents a "textual history" of the Annual Report of the Registrar General from 1839 to 1920, and describes the changing form and content of this seminal publication of public health statistics. He is generally in accord with the account of the evolution of this publication offered elsewhere by Simon Szreter, but while the latter explains the highs and lows of publication on the basis of large-scale changes in the nature of public health institutions, Higgs focuses on the managerial qualities of leading individuals and institutional policies.

There has always been a tension in medicine between skill (the "art") and impersonal quantification. John Senior shows nicely that this emerged even in so technical a field as electrotherapy. Debate within the field pitted advocates of "metrology"—the use of precision instrumentation and standardized dosage—against the defenders of clinical experience as a criterion for determining exact electrical dosage. The essay makes it clear that simple distinctions between technology and clinical experience are insufficient for understanding changing medical practices, for technology can be understood as a simple extension of clinical experience.

Karl Pearson lies at the heart of this book. Eileen Magnello's impressive essay provides an important account of the shift from vital statistics to mathematical statistics. The "bridge" between the two was Francis Galton—but it was Pearson who was the founder of mathematical statistics, introducing such innovations as standard deviation and the chi-square test. His techniques were gradually assimilated into both public health and clinical medicine through figures like Major Greenwood and his student Austin Bradford Hill. J. Rosser Matthews contributes to this story by returning to a subject about which he has already written, the debate between Pearson and Almroth Wright on the opsonic index. The debate nicely illustrates the conflict between modes of reasoning based, in the latter case, on physiologically based clinical research, and, in the former, on new mathematical modes...