Inquisition can go one of two ways. Either it can drive innovation, encourage close inspection and speed progress. Or like the Spanish Inquisition it can stifle, oppress and lead to abuse by controlling powers.
I have discussed earlier how subjects (like pain) differ to objects (like bodies). And that science prefers the concreteness of objects to measure. My CauseHealth talk skimmed over some of the details so I want to explore this further here (see my talk slides here). With pain maybe we have contrived to measure the unmeasurable. So we use a proxy. And what better to measure than numbers. Hey presto! We have the NRS (Numeric Rating Scale: 0-10). We nudge the subject towards the object by giving it an object characteristic. A weird sort of ‘subjectopomorphism’ (derived from its better known cousin anthropomorphism!).
At this point other than being a bit odd I have no qualms. If people want to try to measure pain it might be as good as we’ve got. There are potential problems though. Not only can we start to add qualities that don’t exist. But we can also subtract those that do. At the same time it is easy to lose our grip on metaphysical actuality (reality!). We can lose sight of the change made. That pain becomes something other……. maybe ‘painmark1’. Then we get greedy. Small changes feel reasonable. Ideally we would like pain to be linear and measurable rather than just arbitrary. This helps statistical analysis when the data becomes ordinal. It is more powerful, comparable, accurate and can be used in favoured t-tests, ANOVAs. So there is some motivation to do this. This however requires a firther shift from subject towards object. Another step removed from what we are trying to measure. No longer ‘painmark1’ but ‘painmark2’. The change to VAS assumes that pain can be measured with millimetre accuracy. That there is equal distance between 1 point change in pain from 1-2/10 as 7-8/10.
Kersten et al 2014 set up a study to investigate this. They took people awaiting single joint replacement (knee 60%, hip 40%) who were determined to have relatively stable pain. N=221. Mean age 67yrs. They got them to complete a VAS scale on 7 consecutive days, then weekly over 6 weeks. The data was then analysed through a Rasch model. Georg Rasch developed a probabilistic statistical model to determine fit of interval data (amongst other things). Rasch’s approach to social measurement was more concerned with the individual rather than population, he did not want population assumptions alone to drive measurement techniques. He also recognised that for something to be quantitative and measurable is a scientific hypothesis. Not a given.
The philosopher of science Thomas Kuhn was interested in quantifying and measuring although more in the physical sciences. Even here he found significant difficulties in measurement and the role of quantifying (of which he still regarded central to the physical sciences). He felt measurement could not be separated from theory. In fact the theory would determine the what, where, when, how questions about measuring data. Without this he suggested we were aimlessly wasting time. He saw the quantitative data more for identifying anomaly in theory and noted that ‘reasonable’ agreement between theory and data was commonplace in the physical sciences with much variation between specific sciences. Measurement wasn’t a simple application but often required inquisitive and innovative ideas, technologies, and applications over a period of time.
Take for example temperature in thermodynamics. Measuring in Fahrenheit and Celsius did not just casually happen. In fact scientists had to invent ways in which temperature might be measured. Initially at the start of the 17th Century Galileo (amongst other scientists) developed thermoscopes which used expansion/contraction of gas to move a water level to indicate temperature. At this point no measure or scale. The first diagram appeared in 1617 and by 1638 a rudimentary scale was in place making it a thermometer of sorts. The early forms showed an 8 point scale. In 1654 thermometers became sealed as it was noted air pressure would affect the measurement and they were part filled with alcohol instead of water. In 1665 Huygens proposed standardising measurement around the boiling and freezing points of water. In 1694 Renaldini proposed fixed points on a universal scale. Newton in 1701 proposed 12 point scale between freezing point of water and body temperature. Only in 1724 did Daniel Fahrenheit create a widely used universal scale using mercury (and this was mainly because he was a thermometer manufacturer!) It was 1742 before Anders Celsius proposed a 0-100 scale between ice melting and water boiling points. It took over 100yrs to achieve this measurement progress. The measurement informed by the theory. Still 2 centuries later the Kelvin was proposed as a unit of measurement when theory had indictated the possibility of absolute zero (where thermal motion ceases).
So back to pain. Theory dictates our measurement and at this point it is hard to justify an interval based measurement for something so poorly described. It presents unique difficulties compared to physical concepts such as temperature measurement. We need to decide whether it is possible to measure. Be inquisitive about it’s nature to inform this and don’t expect it to happen overnight! The Kersten study goes on to suggest that pain is not linear in nature. This provides problems for MCID (minimal clinically important difference) which is claimed to be 13-28mm as this would mean different things depending on where the person was on the scale (Bird and Dickson 2001). The scale was also insensitive to change. Once 15 ‘outliers’ were removed VAS did show good internal vailidity. This however doesn’t mean it is testing what it thinks it is. More that it was not affected much by confounding variables.
Maybe as a physiotherapy profession we are natural do-ers. We like physical things, we want to get stuff done. Perhaps as a profession we brush over conceptual development becoming impatient to measure now, get results now, make changes now! Which is why I am excited to see work from Matt Low on concept analysis (amongst others).
Use more inquisition. Use less Spanish Inquisition.
Be more human. Be less robot.
Thanks for reading this far!