Underpredicting Pain: An Experimental Investigation Into the Benefits and Risks

Kaya J. Peerdeman; Andrew L. Geers; Delia Della Porta; Dieuwke S. Veldhuijzen; Irving Kirsch


Pain. 2021;162(7):2024-2035. 

In This Article

Abstract and Introduction


Expectancies can shape pain and other experiences. Generally, experiences change in the direction of what is expected (ie, assimilation effects), as seen with placebo effects. However, in case of large expectation–experience discrepancies, experiences might change away from what is expected (ie, contrast effects). Previous research has demonstrated contrast effects on various outcomes, but not pain. We investigated the effects of strong underpredictions of pain on experienced pain intensity. In addition, we assessed related outcomes including (certainty of) expectations, fear of pain, pain unpleasantness, autonomic responses, and trust. Healthy participants (study 1: n = 81 and study 2: n = 123) received verbal suggestions that subsequent heat stimuli would be moderately or highly painful (correct prediction), mildly painful (medium underprediction; study 2 only), or nonpainful (strong underprediction). Both studies showed that participants experienced less intense pain upon strong underprediction than upon correct prediction (ie, assimilation). Expected pain, fear of pain, and pain unpleasantness were generally also lowered. However, strong underprediction simultaneously lowered certainty of expectations and trust in the experimenter. Study 2 indicated that the effects of strong underprediction vs medium underprediction generally did not differ. Moreover, study 2 provided some indications for reduced heart rate and skin conductance levels but increased skin conductance responses upon strong underprediction. In conclusion, even strong underpredictions of pain can reduce pain (ie, cause assimilation), although not significantly more than medium underpredictions. However, strong underpredictions can cause uncertainty and undermine trust. These findings suggest that healthcare providers may wish to be cautious with providing overly positive information about painful medical procedures.


Expectancies have repeatedly been found to influence pain experiences and are considered to be a core mechanism of placebo and nocebo effects.[5,14,26,34,35] In line with response expectancy theory,[34,35] predictive coding models,[10] and other theoretical accounts,[8,44,62] pain experiences are generally found to change in the direction of what one expects, referred to as assimilation effects,[48–50,59] especially when expectations are certain.[10,27] As such, shaping positive expectancies may improve pain treatment.[16] Under certain circumstances, however, expectancies might have counteractive effects. Most notably, the affective expectation model,[62] as other models,[8,32,44,45] suggests that experiences can change away from expectations, referred to as contrast effects. Contrast effects have been observed in various research fields, particularly when there is a large discrepancy between expectations and experiences.[1,7,11,18,19,36,44,45] However, to the best of our knowledge, currently, no empirical evidence exists for contrast effects on pain. This may be partly due to the use of relatively small and hence unnoticed expectation–experience discrepancies in previous research.

Alternatively, the ambiguous nature of pain experiences may inherently favor assimilation, even when expectation violations are large, but there might be limits to this. In line with earlier research,[36] recent experiments demonstrated a tipping point at which large discrepancies between predicted and actual pain intensity led to less assimilation than moderate discrepancies.[25] Moreover, there are some indications that even when underprediction of pain leads to assimilation, negative effects in the form of increased pain intensity expectations, expectation uncertainty, autonomic responses, and task disruptions can occur.[2,3] Observational and qualitative research further indicates that patients commonly hold overly positive expectations of medical treatments, and violation of these expectations has been related to disappointment, distress, and treatment dissatisfaction.[13,23,29,38] Expectation violations might also undermine trust in the caregiver, which is core in health care[22] and placebo effects.[16] Further research is required to obtain a comprehensive understanding of the benefits and risks of large expectation–experience discrepancies on pain, particularly when expectancies are overly positive. This will provide essential insights into how painful medical procedures and analgesic treatments can best be introduced to patients.

In 2 preregistered experiments, we investigated the effects of strong underpredictions of pain. Our primary hypothesis was that a suggestion of no pain from a moderately or highly painful heat stimulus (strong underprediction) would increase experienced pain intensity (ie, elicit a contrast effect) as compared to a correct prediction. In study 2, we also compared this with a medium underprediction of pain to assess possibly reduced assimilation upon large vs medium expectation–experience discrepancies. Secondarily, we hypothesized that effects on pain expectations upon undergoing the expectation–experience (mis) match would be comparable with effects on pain intensity as experiences shape expectations.[3,10] We additionally explored effects on expected and experienced pain during multiple trials, including upon repetition of the suggestions, as well as effects on expectation certainty, fear of pain, pain unpleasantness, heart rate, and skin conductance. Finally, we explored additional psychological outcomes (eg, trust and discrepancy awareness) and possible moderation by individual characteristics (eg, optimism).