Institute for the Study of
Healthcare Organizations & Transactions



Laboratory Research in the Modern Era

For the most part, however, for 100 years after Elliotson (1843) and Esdaile (1846), interest in hypnosis and pain was largely academic, and confined to experimental work. Robert Sears (Sears, 1932) conducted a pioneering study of hypnotically induced pain analgesia, while Frank Pattie (Pattie, 1937) did the same for tactile anesthesia. They and others (e.g., Brown & Vogel, 1938; West, Niell, & Hardy, 1952) showed clearly that hypnotic suggestions altered conscious perception of the pain or touch stimuli, if not always of involuntary or indirect responses as well -- a topic to which I will return later. In the late 1950s and early 1960s, when experimental research on hypnosis was revived by Martin Orne, Ernest Hilgard, J.P. Sutcliffe, Theodore Barber, and their associates, hypnotic analgesia was there in the center of things (e.g., Barber, 1963; Barber & Hahn, 1962; Shor, 1962; Sutcliffe, 1961).

The attraction of analgesia to the first generations of modern hypnosis researchers was, I think, threefold. (1) With the possible exception of hypnotic amnesia (which after all gave hypnosis its very name), analgesia is the most dramatic of the alterations in consciousness observed in hypnosis. (2) It is also the most susceptible to "objective" measurement -- although, as I will indicate later, this feature has been somewhat overstated. (3) And, of all the phenomena of hypnosis, it has the most potential for practical use. Therefore, it is not surprising that when Jack Hilgard began his systematic study of hypnosis, he would quickly come to focus on analgesia.

The single most important fact about hypnosis is that there are individual differences in response to hypnotic suggestion. Unfortunately, hypnotizability cannot be predicted by the usual sorts of paper-and-pencil questionnaires, but must be assessed directly by means of performance-based assessments of  hy_pa09.gif (9033 bytes)hypnotizability analogous to intelligence tests. The Stanford Hypnotic Susceptibility Scales, devised by Hilgard and Andre Weitzenhoffer in the late 1950s and early 1960s, begin with a standardized hypnotic induction procedure followed by a series of suggestions for imaginative experiences. The subject’s response to each of these suggestions is scored according to an objective behavioral criterion. When these  scales are administered to unselected samples, they yield a roughly normal distribution of scores: while relatively few subjects are entirely refractory to hypnosis, what Hilgard referred to as "hypnotic virtuosos" comprise only about 10%-15% of the population.  Click on the image to view an enlarged version.

The relevance of hypnotizability to hypnotic analgesia is shown clearly in an early laboratory study by Hilgard involving cold-pressor pain stimulation in which subjects’ hands and forearms are immersed in circulating ice water (Hilgard, 1967; Hilgard, 1969, 1978). The result is a very good laboratory analog of clinical pain.  Under normal conditions, pain mounts rapidly over the course of a minute hy_pa10.gif (6815 bytes)or so, as indicated by subjects’ reports on an open-ended scale. However, when hypnotized and given suggestions for analgesia, subjects of relatively high hypnotizability gain substantial pain relief: they may still feel some pain, but even after 60 seconds the majority find the levels to be tolerable. Click on the image to view an enlarged version.

Of course, pain has two components, measured on the McGill Pain Questionnaire as sensory pain, providing information about the location and severity of an injury, and suffering, or the "meaning" of the sensory pain.  During his psychophysical hy_pa11.gif (7501 bytes)investigations Hilgard found that hypnosis alters both of these components. In one study, involving ischemic muscle pain produced by cutting off the flow of blood to the forearm, both pain and suffering were reduced essentially to zero in a group of highly hypnotizable subjects (Knox, Morgan, & Hilgard, 1974). A recent brain-imaging study by Rainville and his associates showed that strategically worded suggestions can dissociate the two components of pain, selectively altering one but not the other (Rainville, Duncan, Price, Carrier, & Bushnell, 1997).  These hy_pa12.gif (21992 bytes)investigators then used this dissociation in a provocative brain-imaging study which indicated that the two components of pain have different biological substrates: sensory pain in the primary somatosensory cortex, and suffering in the anterior cingulate cortex -- the same place where other investigators have recently located romantic love!  (The figure shows activation in the right hemisphere because the pain stimulus was applied to the left hand.) Click on the images to view enlarged versions.

Along the same lines, an unbelievably spartan study by John Stern and his associates compared the effectiveness of seven different challenging agents -- hypnosis,  acupuncture, placebo acupuncture, morphine, aspirin, diazepam, and hy_pa13.gif (8937 bytes)placebo pill -- against two kinds of pain in a within-subjects design that entailed a total of 280 pain trials per subject (Stern, Brown, Ulett, & Sletten, 1977). Analysis of the subjects’ cold-pressor pain ratings showed that hypnosis was superior to all otherhy_pa14.gif (7302 bytes) challenging agents, especially for those subjects who were highly hypnotizable. Interestingly, only hypnotic analgesia was correlated with hypnotizability. Similar findings were obtained for ischemic muscle pain. Click on the images to view enlarged versions.

More recently, a meta-analysis of clinical and experimental studies reported by Montgomery and his colleagues compared hypnosis against a number of other cognitive-behavioral interventions (Montgomery, DuHamel, & Redd, 2000).  The hy_pa15.gif (6914 bytes)studies in question were a mix of the clinical and experimental, with most of the experimental studies involving cold-pressor pain and the clinical studies including patients suffering from both burns and cancer. Comparing hypnosis to a no-treatment control, Montgomery et al. found that the effect of hypnosis corresponded to a standardized effect size of d = .74, which counts as a moderately large effect size by conventional standards. More than a quarter of the comparisons yielded very large ds of 1.0 or more.  The effect was bigger in clinical patients than in experimental subjects,  hy_pa16.gif (7161 bytes)and especially large among those of high hypnotizability.hy_pa17.gif (7338 bytes) Comparing hypnosis to a variety of other cognitive-behavioral treatments tested in the same studies, Montgomery et al. found no particular advantage for hypnosis, but this analysis did not take hypnotizability into account.  It is likely that hypnotizable subjects would have gained more from hypnosis than from any other psychological treatment. Click on the images to view enlarged versions.

Click on the next section, Clinical Studies of Hypnotic Analgesia

Or, click on to read other sections:  

Prelude to the Modern Era

Mechanisms of Hypnotic Analgesia

Efficacy of Hypnotic Analgesia



John F. Kihlstrom, PhD

Plenary address presented at the annual meeting of the American Pain Society, Atlanta, Georgia, November 3, 2000.  The point of view represented in this paper is based on research supported by Grant #MH-35856 from the National Institute of Mental Health.  I thank Lucy Canter Kihlstrom for her comments.  Painting by Richard Bergh (1887).


Copyright © 2000 Institute for the Study of Healthcare Organizations & Transactions

Last modified:  04.08.2010 02:58 PM