From the days of my infancy as a physiotherapist, I was raised on the teaching that pain should be measured by subjective means only. You have to ask the patient. You cannot presume to judge the magnitude of their experience: you ask them, and they tell you. If they say it’s a 2 out of 10, that’s what it is. If they say it’s an 11 out of 10 (I see you smile as you recognise this experience), that’s what it is. Countless students have challenged this: “What if the patient walks, jogs and hops on the painful ankle, and then says the pain is 10/10?” But the reply was always: “Their pain is as bad as they say it is. You have no better way of measuring it.”
I think that this is standard fare in the clinical world, for many good reasons. But take a look at the psychology literature, and you’ll see experimental “pain” being measured using all sorts of other approaches. Subjective ratings are not enough: one must also measure the psychophysiological expressions of pain. Arousal is a big focus for measurement: skin conductance is used to estimate how much the participant is sweating, or the acoustic startle response – the strength of the reflexive eyeblink in response to a sudden noise – is used to estimate how alert the participant is. Avoidance behaviour can also be assessed: response latency time tells us how long the participant hesitates for before they respond to a given instruction. But can these really tell us about pain?
As a clinical physiotherapist, my automatic response is “No! Those are behavioural responses! They may be related to pain, but they don’t quantify pain itself.” And that’s all well and good until someone points out that subjective pain ratings, too, are behavioural responses. We have no way of knowing how accurately that subjective pain rating corresponds with the actual, lived experience of the pain.
In his doctoral thesis, Geert Crombez provides an accessible explanation of why the psychology literature displays so many different proxy measurements for pain. He argues that the actual, lived experience of pain cannot ever be directly measured. It is a latent construct. The only way to quantify something latent is by using proxy measurements to quantify things we think are related to the latent construct, and then infer the presence, absence or quantity of the latent construct. Crombez points out (and I paraphrase) that, in the context of wider society, humans infer the presence or absence of pain in others based on what we observe. We observe behaviour: she sat down very suddenly and grabbed her knee, pressing it hard and bending over it with her eyes closed and a grimace on her face. We observe physiological changes: beads of sweat started to form on her forehead. We observe an antecedent stimulus: she had just been walking very fast, looking over her shoulder, when she walked into that bollard and stopped suddenly. And we observe further behaviour: she said, “Ahhh, it hurts.” The verbal communication of pain is just another behaviour.
On the basis of this, Crombez suggests that, in research, pain ought to be inferred on the basis of 3 kinds of measurement:
1) Subjective reporting
2) Psychophysiological measures (muscle activity, sweating, etc.)
3) Overt behaviour (she clutched her knee and grimaced)
I like this suggestion. The idea that we could use more than one measure, so that what we find in one might be backed up by what we find in the other, feels reassuring. And it makes sense to me that pain is not directly measurable. While I don’t think that pain can be inferred in the absence of subjective reporting, I can see the benefit of substantiating the subjective report with other measures within a research context. This approach also reflects how humans tend to infer the existence of pain in natural, social contexts.
But there are some problems, too. The correlation between the measures of these three aspects is not always as good as one might expect. In fact, with psychophysical measures, the relationship between a chosen measurement and the construct it is intended to measure may also be poor. Now, I am no psychologist, but it seems to me that this raises a question: are we sure that we know what we are measuring?
Knowing what we are measuring can be a problem when observing overt behaviour, too – in both clinical and research settings. When a person grimaces are they expressing pain? Or anticipation of pain? When they hesitate to do something are they scared? Reluctant? Tired and bored? Uncertain of the instruction? Distracted?
These are not new questions. Measurement of pain has been debated in many ways, and in many spheres. But this blog site brings together many people from many different backgrounds, which provides an opportunity for us to have a robust and helpful discussion about this. Whether you are a clinician or a researcher, I invite you to give this some thought and use the comments section: let’s talk about it.
What do you reckon? Which measures do you consider reasonable bases for inferring that someone is in pain or for quantifying their pain? What would you urge us to consider more carefully as we discuss this quandary?
Tory arrived from South Africa to start her PhD at BiM. She is a physiotherapist who worked clinically before turning her focus toward research. She is interested in pretty much anything related to pain and neuroscience, thanks to some particularly inspirational teaching by Romy Parker during her undergraduate training at the University of Cape Town.
Tory is rapidly developing a fondness for Australia’s amazing TimTams, but is rather worried about whether she will survive the Adelaide winter. She loves sports but has been forced to take some time out due to a knee injury. She has used this extra time to try her hand at table tennis, with which she is making moderate progress, and rowing, at which she should not yet be left unsupervised…
Crombez G (1994) Pijnmodulatie door anticipatie (Pain modulation through anticipation). Doctoral dissertation. University of Leuven (Belgium).