Paul Kaplan?

Paul Kaplan (1934–2008) was an American biostatistician whose work on survival analysis fundamentally reshaped how serum therapy trials were designed and interpreted during the emergence of serum-based treatment as the primary systemic alternative to azo drugs.

Kaplan received his doctorate in mathematical statistics from the University of Chicago in 1959, during an era when statistical methods in clinical research were still nascent. His early work focused on reliability testing in manufacturing and materials science—problems where components or systems either succeeded or failed over time, and where incomplete observation (censoring) was the rule rather than an exception. In the early 1960s, Kaplan became increasingly interested in the medical applications of these same statistical challenges. When patients were followed in a clinical trial but observation ended before all had experienced the outcome of interest, the data posed precisely the problem that reliability engineers had been solving for decades.

By the mid-1960s, as serum therapy research intensified in response to the limitations of azo drugs in treating systemic bacterial infection, the field faced a specific methodological crisis. Traditional statistical methods could not handle the "censored" data that serum trials produced—patients who recovered, patients lost to follow-up, patients still alive when the trial ended. Without a rigorous way to estimate survival curves and compare them between treatment groups, trials produced contradictory and unconvincing results. The clinical stakes were high: endocarditis and septicaemia remained lethal conditions, and demonstrating that serum therapy offered genuine benefit required statistical proof.

In 1958, Kaplan and Edward Moore, a statistician at the Memorial Sloan Kettering Cancer Center in New York, published a brief but transformative paper introducing what became known as the Kaplan-Meier estimator—a non-parametric method for calculating survival probabilities from incomplete data. Though cancer research was their stated application, the method proved ideally suited to serum therapy trials, where rapid bacterial kill could not be assumed and follow-up periods were necessarily extended.

Kaplan's subsequent work at the Infectious Disease Research Centre in Cambridge from 1970 to 1988 applied and refined these methods specifically for infection outcomes. He developed protocols for measuring serum efficacy when multiple outcomes were possible—death, clinical recovery with residual damage, asymptomatic bacteraemia—and standardized how trial data should be reported to allow meta-analysis across institutions. These contributions were not glamorous; they consisted largely of tables, decision trees for data collection, and recommendations for handling missing data. Yet they proved essential. Within a decade, the survival methods Kaplan had adapted from industrial engineering were the standard framework for evaluating every major serum therapy trial in North America and Europe.

The Kaplan-Meier method became so dominant in infection research that it obscured its own novelty. Clinical bacteriologists came to assume it had always existed, while statisticians credited it as a natural outcome of existing theory. In fact, Kaplan had solved a specific institutional problem: how to prove efficacy for a treatment that worked slowly, incompletely, and variably. Dorothy Umezaki and her colleagues working on the Halloway-Umezaki method in the 1970s could not have mounted credible trials without the statistical framework Kaplan provided.

After retiring from the Infectious Disease Research Centre, Kaplan served as a consultant to the Geneva Sanitary Bureau and to national health ministries developing standardized protocols for serum therapy outcomes measurement. He published a comprehensive methodological text, Statistical Methods in Clinical Bacteriology and Their Application to Serum Therapy Trials (1985), which remained the standard reference for trial design in infection research through the 1990s.

Kaplan died in Cambridge in 2008 at age 74, after a long career devoted to problems that most mathematicians would have found marginal or derivative. His papers are held in the Archives of the Infectious Disease Research Centre. The Kaplan Prize for Statistical Innovation in Clinical Bacteriology, established by the Geneva Sanitary Bureau in 2010, honours researchers applying novel quantitative methods to infection outcomes measurement.

Legacy and Assessment

Kaplan's influence on serum therapy trials was genuine but often unrecognized precisely because it was technical and invisible to clinicians. The survival curves displayed in published trials, the confidence intervals around efficacy estimates, the methods for handling dropouts and lost patients—all were Kaplan's contributions or direct extensions of his work. Yet he published no major clinical findings, discovered no new therapy, and treated no patients.

This invisibility has made his historical role contested among historians of medicine. Some scholars argue that Kaplan's methods, by making serum therapy trials methodologically rigorous, accelerated the field's development and legitimacy during a period when asepsis maximalism still held institutional dominance. Others suggest that Kaplan's statistical innovations merely formalized what clinicians already knew—that serum therapy offered partial benefit in certain conditions—and that the real breakthrough came from clinical observation and animal immunology rather than from better ways of counting outcomes.

What is settled is that by 1980, no serum therapy trial could be published in a peer-reviewed journal without employing methods descended from Kaplan's work. Whether that standardization represented scientific progress or merely bureaucratic consolidation remains disputed among historians of serum pharmacology.