Publications
APS Bulletin • Volume 9, Number 1, January/February 1999
History of Pain
Marcia Meldrum, PhD, Department Editor
Each Patient His Own Control: James Hardy and Henry Beecher on the Problem of Pain Measurement
Marcia L. Meldrum, PhD
DEPARTMENT EDITOR'S NOTE: In the years after the Second World War, a small number of physicians and scientists were engaged in debate about the problem of pain, which brought into focus many of the issues familiar to those in the field today. One of those debates, between the physiologist James D. Hardy and the anesthesiologist Henry Knowles Beecher, concerned the problem of pain measurement.
From 1929, the National Research Council began sponsoring a systematic search for an effective nonnarcotic analgesic, which by the early 1950s had led to the isolation of some 500 possible compounds (Eddy, 1956). The question was, How were they to be evaluated for efficacy in comparison with each other and with standard drugs such as morphine and aspirin? The method proposed by James Hardy of Cornell University and his associates, Harold G. Wolff and Helen Goodell, was to quantify the pain threshold as a function of stimulus intensity. The Cornell team carried out a long series of experiments on pain sensation, discrimination, and reactions from 1950 to 1959. Their preferred method employed radiant heat from a high-wattage lamp, directed at the forehead or hand of an experimental subject, often one of the researchers themselves. This method presented no known risk of harm to the subjects and permitted very precise calibration of stimulus intensity, extent and duration of exposure, and other variables (Hardy, Wolff, & Goodell, 1952; Wolff, Hardy, & Goodell, 1940).
Hardy and his associates developed a scale of pain measurement based on the "jnd," or just noticeable difference, which was the subject's perception of a change in pain intensity. A jnd can be defined quantitatively in terms of stimulus magnitude, and one "dol" is defined as being equivalent to two jnds. Each subject's threshold and maximum tolerance were measured in dols. After receiving an initial exposure to a stimulus to determine a baseline threshold, the subject received a specific dose of a particular drug and went through a series of exposures. With effective analgesia, the patient's pain threshold occurred at a stimulus level above the baseline; the investigator reactivated the stimulus and charted the threshold at set time intervals until the baseline was again reached to determine the duration of analgesic effect. Wolff et al. (1940, 1941) produced a series of graphs showing the relative duration and extent of change in threshold of all established analgesics (Hardy, 1956).
FIGURE 1. Hardy-Wolff-Goodell Time-Action Curves for Acetylsalicylic Acid at Different Dosage Levels.
NOTE. Each dosage level point represents the average threshold level of 2-3 subjects. Source: Wolff, Hardy & Goodell (1941), p. 64. Reproduced from the Journal of Clinical Investigation, 20 (January) by copyright permission of the American Society for Clinical Investigation.
Despite the Hardy team's meticulous approach, other investigators found their results difficult to replicate. Robert A. Kuhn and R.B. Bromiley at Johns Hopkins reported a great deal of variation in threshold change among subjects receiving the same dosage of morphine. One subject's threshold actually fell slightly, while two others reported no perception of pain until their skin had begun to burn. Kuhn and Bromiley and other researchers who attempted to replicate the Cornell experiments with aspirin reported no threshold change at all with this commonly used analgesic (Kuhn & Bromiley, 1951; Sonnenschein & Ivy, 1949; Whyte, 1951).
The major challenge to Hardy, Wolff, and Goodell came from Henry Beecher and his colleagues at the Anesthesia Research Laboratory at the Harvard Medical School. Beecher's ideas had been greatly influenced by his famous observations of wounded soldiers on the Anzio beachhead who rarely complained of pain (Beecher 1946). He theorized that the perception of pain integrated physical sensation with the "reaction component," which depended on a number of variables, including age, gender, ethnicity, experience, fatigue, anxiety, distraction, suggestion, the external environment, and the time of day. Because of this reaction component, he wrote, the pain experience was "complex, subjective and different for each individual" (Beecher, 1957, p. 165).
Beecher's concept of the reaction component became the basis of an extended critique of the experimental measurement of pain and of the Hardy, Wolff, and Goodell method in particular, which he reiterated at many professional meetings in the 1950s, in a long review article for Pharmaceutical Reviews in 1957, and in his major work, The Measurement of Subjective Response in 1959. He presented two arguments based on these issues: first, the noncomparability of the experimental and clinical environments and, second, the dominance of the reaction component.
The Hardy group and their subjects were very different people from the wounded soldiers or from pain patients. They had no fear or anticipation of serious injury; instead, they expected and were prepared to tolerate pain as part of an interesting and potentially important research project. The Hardy team's reports showed that with each new subject, it took from four to eight sets of exposures to establish the baseline threshold. After these several repetitions, the subjects became experienced and were readily able to assist in charting precise and replicable changes in threshold. Beecher (1956) wrote "their beautiful curves perhaps oversold their method" (p. 13), suggesting a spontaneous sequential series of discrete events. In fact, he suggested, the neat graphs represented the educated expectations of pain of an investigator and a subject (Beecher, 1957). Hardy and his associates, responding to these criticisms, readily agreed that sensory perception "is but a part of the total pain experience" (Hardy, 1956, p. 42) and that "the feeling state may perhaps be the most relevant aspect of pain to the one who suffers" (Hardy, Wolff, & Goodell, 1952, p. 24). They also recognized that "important bodily reactions below the level of conscious activity, may contribute new sources of noxious stimulation and pain" (p. 24). But, as pain sensation was the only aspect readily accessible to quantitative measurement, they had chosen to give "secondary consideration" in their work to "associated phenomena" (p. 24). Their studies were designed to provide an objective measure of analgesic effectiveness to aid clinicians when they were prescribing analgesics.
To develop this objective tool, Hardy, Wolff, and Goodell had set up a controlled situation to create a "pure pain" experience dissociated from reaction and consisting only of sensation. For their subjects, "the attitude toward the pain was one of relative relaxation, interest and satisfaction... There was complete freedom from anxiety. Very painful procedures could be tolerated with immunity" (Wolff, Hardy, & Goodell, 1940, pp. 659, 677).
FIGURE 2. Analgesic Potency of Morphone at Different Dosage Levels When Treated as Unknown.
SOURCE. Keats, Beecher, & Mosteller (1950), p. 42. Reproduced by permission of The American Physiological Society.
Beecher commented dryly that "for a pure pain sensation this is quite a description of quite a reaction" (Beecher, 1953, p 4). In his view, the subjects' tranquility and interest, although quite unlike the anxiety and suffering of pain patients, constituted an emotional state that affected their perception of pain and indicated that the reaction component was still present.
Beecher refused to consider the comparisons of analgesic effectiveness produced by the Cornell group as applicable to the relief of clinical pain, so how, then, did he propose to compare and evaluate analgesic compounds in a clinical setting? Beecher and his colleagues (Jane Denton, Arthur Keats, Louis Lasagna, and Frederick Mosteller) described their proposed method in a series of papers published from 1949 to 1954. They conducted a series of randomized and blinded trials of various new analgesics, standard drugs, and a placebo on postoperative patients with varied diagnoses. Because each patient's pain experience was individual and noncomparable to that of others in the same group, the Harvard team advocated the use of a crossover design. Each patient received two or more drugs in a randomized sequence, so that the patient "served as his [or her] own control" and the analgesics were tested against "the same order of pain" (Keats, Beecher, & Mosteller, 1950, p. 35). Since they considered subjective descriptions of pain relief ("slight," "moderate," "complete") to be difficult to quantify, Beecher and his group employed a simple binary choice.
They defined the term relief as a 50% reduction in pain, "a judgment patients have found easy to make" (Beecher, 1957, pp. 94-95; Beecher, Keats, Mosteller, & Lasagna, 1953; Keats, 1950, p. 35). By comparing graduated dosages of morphine against an unknown (actually 10 mg of morphine), they were able to demonstrate the discriminatory clarity of their method (Denton & Beecher, 1949; Keats et al., 1950, p. 42).
This design, which theoretically held the unknown variables affecting pain perception constant between each drug administration, produced a measure, called "percentage of patients achieving relief," that quantified the analgesic effectiveness of a drug, while the patient's subjective experience remained amorphous and opaque. The crossover method was replicated by Raymond Houde at Sloan-Kettering and by many other investigators and had become the standard for analgesic evaluations by 1959. The pain field has since developed many new tools and measures to help clinicians assess patients' verbal reports of their subjective experiences.
FIGURE 3. Hardy, Wolff, and Goodell's Experimental Setup.
SOURCE. Hardy, Wolff, & Goodell (1952), p. 269. Reproduced by permission of Lippincott Williams and Wilkins.
The Hardy-Beecher debate, however, helped to crystallize the problem of pain in the United States during the early postwar years. Beecher's arguments were aimed not simply at the Cornell team, but at the experimental study of pain in animals and under controlled laboratory conditions. With his statement that pain is "different for each individual," he asserted the clinical claim that only the physician who deals directly with the individual patient can learn to understand the subjective phenomenon of pain. But his insistence on individuality, as Hardy pointed out, complicated the development of objective tools for general use.
Even the percentage of patients reporting 50% relief is meaningless if the term "50% relief" has a different meaning for each patient. By characterizing the dilemma, Hardy and Beecher helped to address the pain problem in a new way.
Hardy's efforts to calibrate pain to a quantitative scale and Beecher's emphasis on the importance of the reaction component laid a foundation for much creative work in pain measurement. The McGill Pain Questionnaire, the Differential Descriptor Scale, and the Slide Algometer, for example, offer workable tools that differentiate between the sensory and affective experience of pain, but each pain patient presents a new assessment problem to clinicians.
By defining the dilemma, Hardy and Beecher helped to "problematize" pain in a new way, which still reverberates in the field.
References
Beecher, H.K. (1946). Pain in men wounded in battle. Annals of Surgery, 123(1), 95-105.
Beecher, H.K. (1953). A critique of Hardy, Wolff, and Goodell's papers. Unpublished manuscript, Henry Knowles Beecher Papers, Countway Library of Medicine, Harvard University School of Medicine.
Beecher, H.K. (1956). Limiting factors in experimental pain. Journal of Chronic Diseases, 4(1), 11-19.
Beecher, H.K. (1957). The measurement of pain: Prototype for the quantitative study of subjective responses. Pharmacological Reviews, 9(1), 59-209.
Beecher, H.K., Keats, A.S., Mosteller, F.C., & Lasagna, L. (1953). The effectiveness of oral analgesics (morphine, codeine, and acetylsalicylic acid) and the problem of placebo "reactors" and "non-reactors." Journal of Pharmacology and Experimental Therapeutics, 9(1), 393-400.
Denton, J. & Beecher, H.K. (1949). New analgesics. Journal of the American Medical Association, 141, 1051-1057, 1146-1153.
Eddy, N.B. (1956). The search for new analgesics. Journal of Chronic Diseases, 4(1), 59-71.
Hardy, J.D. (1956). The nature of pain. Journal of Chronic Diseases, 4(1), 22-51.
Hardy, J.D., Wolff, H.G., & Goodell, H. (1940). Studies on pain: A new method for measuring pain threshold-Observations on spatial summation of pain. Journal of Clinical Investigation, 19(4), 649-657.
Hardy, J.D., Wolff, H.G., & Goodell, H. (1952). Pain sensations and reactions. Baltimore: Williams and Wilkins.
Keats, A.S., Beecher, H.K., & Mosteller, F.C. (1950). Measurement of pathological pain in distinction to experimental pain. Journal of Applied Physiology, 3(1), 35-44.
Kuhn R.A., & Bromiley, R.B. (1951). Human pain thresholds determined by the radiant heat technique and the effect upon them of acetylsalicylic acid, morphine sulfate, and sodium phenobarbitol. Journal of Pharmacology and Experimental Therapeutics, 101(1), 47-55.
Sonnenschein, R.R., & Ivy, A.C. (1949). Failure of antipyretic drugs to alter normal human pain thresholds. Journal of Pharmacology and Experimental Therapeutics, 97, 308-313.
Whyte, H.M. (1951). The effect of morphine and aspirin on heat pain. Clinical Science, 10(3), 333-345.
Wolff, H.G., Hardy, J.D., & Goodell, H. (1940). Studies on pain: Measurement of the effect of morphine, codeine, and other opiates on the pain threshold and an analysis of their relation to the pain experience. Journal of Clinical Investigation, 19(4), 659-677.
Wolff, H.G., Hardy, J.D., & Goodell, H. (1941). Measurement of the effect on the pain threshold of acetylsalicylic acid, acetanilid, acetophenetidin, aminopyrine, ethyl alcohol, trichloroethylene, a barbiturate, quinine, ergotamine tartrate and caffeine: An analysis of their relation to the pain experience. Journal of Clinical Investigation, 20, 63-65.
Marcia Meldrum is a DeWitt Stetten Memorial Fellow at the National Institutes of Health in Bethesda, MD.