In 1940, a newly introduced pain measuring apparatus heralded a revolution in analgesic testing. Hailed as ingenious and elegant, the dolorimetric method elaborated by James Hardy, Harold G. Wolff and Helen Goodell at Cornell University promised to overcome the failings and fluctuations of human pain perception that had plagued previous attempts at measuring the efficacy of analgesics. The appeal of the dolorimeter lay in its alleged ability to produce easily identifiable, uniform and quantifiable responses to pain. But in other experimenters’ hands, the method often failed to reproduce the results reported by Hardy, Wolff and Goodell. By the early 1950s, it had become evident that the dolorimeter would not play the major role it had been forecast in the evaluation of analgesic drugs in the United States. How can we explain this transformation?

The failure of the apparatus to produce consistent results does not fully explain why the Hardy-Wolff-Goodell method was largely abandoned as a gauge of analgesic efficacy. The first part of my paper will describe how attempts at solving problems in the diffusion of the dolorimeter gave way to a more fundamental debate about what constituted legitimate means of standardizing experimental conditions. The second will demonstrate that the rejection of the Hardy-Wolff-Goodell apparatus, and the adoption of clinical trial methodologies, can be linked to shifts that took place between 1940 and 1950, in the setting and sponsorship of analgesic testing.

Two principal arguments were offered to explain why the dolorimeter did not perform uniformly in different sites. Critics charged that only subjects who had been trained and informed gave predictable responses with the machine. Hardy et al. conceded that "familiarization" with the method, combined with the proper attitude, protected subjects from distracting mental influences. Critics, however, took the non-interchangeability of subjects, and their knowledge, as evidence that they were biased. If predictable results could only be obtained by standardizing and shaping subjects’ minds, then the method was not a reliable indicator of "true" analgesic efficacy. A second reason given for the methods’ inconsistencies was that the experience of pain could not be standardised. This position questioned the authenticity and meaningfulness of the response produced by the instrument. From the mid-1940s, it became increasingly common, particularly for clinical researchers, to state that the "natural" pain of illness and injury differed fundamentally from "artificial" pain induced in the laboratory. The very predictability and measurability of "laboratory pain" made it a poor substitute for "real pain".

Researchers who worked on analgesic testing could emphasise the need for large numbers of untrained, "naturally" suffering subjects because they had obtained access to better funding, more clinical research facilities and abundant patient populations during the course of the 1940s. Various developments facilitated this increased availability of resources: Public health and military interests in synthetic opiates created a demand for methods to compare the analgesic power of newly synthesized drugs with morphine and codeine, for which they were intended to serve as substitutes. A Committee on Drug Addiction combined the grants and facilities provided by these federal agencies with funding it attracted from pharmaceutical companies. The researchers who benefitted from this funding increasingly consisted of academic clinicians, notably anaesthetists, who had access to large populations of hospital patients, and who were allied with pharmacologists and statisticians. Increased access to resources for clinical testing, and the political pressure exerted on academic researchers and the pharmaceutical industry to follow certain rules of experimental procedure, such as placebo-controlling and randomization, favored the adoption of clinical trial methods to measure the relative efficacy of analgesics. The features of laboratory testing methods, like the dolorimeter, had become less valuable.

This paper contributes to an understanding of the re-conceptualization of pain during a pivotal period: the 1940s and 50s. While other historians, notably Marcia Meldrum (Departures from the Design, unpublished PhD thesis) have pointed to shifts in the use of analgesic testing methods and understandings of pain, this paper anchors these transformations more closely to the material practice, conditions and organization of medical research, and particularly to the activities of the Committee on Drug Addiction. This was also an important period in the organization and development of anaesthesia as a clinical research specialization and in the reform of therapeutic evaluation methods, as described by Harry Marks in The Progress of Experiments. I believe that a close examination of the reception of an instrument such as the dolorimeter must be situated within these important and relevant trends. In addition, little attention has been payed to the role played by American anaesthetists in running clinical trials of analgesics. However, my archival research has revealed that some of the foremost American anaesthetists, such as John Adriani, Henry K. Beecher, Arthur Keats, and John J. Bonica were very interested in issues of experimental design for the testing of analgesic drugs. Their participation in testing analgesics was both a sign and an agent of shifts in the field of drug evaluation that affected the reception of the dolorimeter.

The conclusions presented in this paper are based on research I conducted in the archives of various researchers who operated and criticized the Hardy-Wolff-Goodell dolorimeter, including the personal papers of Harold G. Wolff, Janet Travell, Henry K. Beecher and William K. Livingston. The archives of the Committee on Drug Addiction, a committee affiliated with the National Academies of Science and supported by the Public Health Service, military agencies and pharmaceutical companies, also provided revealing information about the fate of the Hardy-Wolff-Goodell dolorimeter within the broader context of analgesic testing at the time. This research is part of a larger project that examines the history of attempts to make pain assessment more objective. My approach in this project is to focus on the social value of technologies of pain assessment. In this case, I identified the social projects that generated demands and support for the creation and implementation of pain assessment technologies, and the social arrangements that were fostered by the use of these methods.