In my opinion, pain is multifactorial and therefore, have a great subjective factor. So, when one try to use a quantitative measure at a qualitative variable, it has to be well analyze because the risk of error is high. On the other hand, the clinical studies need a way of measure the pain phenomenon, so the VAS may be used if we recognize previously its limitation. Also it shall not be the only measures, since it tries to evaluate just the pain intensity.

One of the greatest error of those who deal with pain is to think that the VAS ( or the NRS, SVS, etc) scale measures PAIN. It does not! It measures only one limited aspect or component of the PAIN experience i.e. the intensity of pain. This is why hundreds of articles, peer reviewed by distinguished reviewers, are of low quality. Authors should always stress that when Pain scales like the VAS or the NRS are used, they measure only the mere intensity of pain and not other components of the PAIN experience. Moreover, another bias when reporting VAS/NRS values is the way they are administered: 1. It is crucial to use always the same way of administration i.e. 0 stands for no pain and 10 for the worst intensity of pain I can imagine (not maximum pain or other staff like this); 2. When reporting pain intensity in a sample one should use median scores and use congruent statistics.

The VAS and other single item rating instruments are meant for ease of use in the clinical setting. They provide a baseline assessment of intensity, however clinicians are also supposed to ask about location, duration, onset, alleviating/aggravating factors, etc...These tools are also often not useful unless we understand what the number means to the patient. A 5 for someone in acute pain with no prior severe pain incidences may be truly significant pain whereas a patient with chronic pain, for instance patients with chronic pancreatitis, may live and be happy with a 5 and are only unhappy when the pain reaches a 7. This is especially true given some patients would prefer some level of manageable pain over the side effects of opioids.

Additionally, many older adults do not identify pain as pain, but as soreness, dullness, achy (mainly related to osteoarthritis or compression fractures). Therefore how the scale is presented and what prior information about the patient the clinician brings to the table are helpful for fully understanding the patient's pain. In patients with cognitive impairment, particularly moderate to severe dementia, we have to use behavioral scales such as PAINAD, PACSLAC, DOLOPLUS to name a few.

Finally, in controlled research experiments, we should be using more than just a pain intensity score. Something like the Brief Pain Inventory or McGill Pain Questionnaire as these are multi-dimensional standardized instruments that can help us to identify pain better and understand the various dimensions of patients' pain.

It's funny, when asked about the VAS many pain scientists react negatively and disparagingly and yet the VAS continues to be used in a very high volume of peer reviewed articles (as noted below). There are problems with the VAS and the most important is the (almost laughable) way the VAS reduces pain to a single numeric. Nevertheless, I would defend the continued use of the VAS on three grounds.

First, the VAS is intuitive, quick and simple to administer. As noted below, it is ideal for use in clinical settings but, as it happens, simplicity and quickness are also very useful in research settings. Despite the narrowness of the scale (or maybe because of it) most people quickly grasp what they are being asked and generally provide sensible approximations of their pain experience.

Second, the VAS correlates very nicely with neural activity in regions of the brain generally believed to be involved in pain experience. Those regions encompass both sensory (intensity) and affective (emotional) processing. (Tor Wager and colleagues have suggested that brain imaging may replace VAS scales (and other subjective measures) as a better, more objective, measure of pain. Mark Sullivan, myself, and others have argued against this because pain is inherently subjective. We will be debating this issue at the IASP in October).

Third, the simple 0-10 VAS also correlates well with other subjective measures that capture additional, multi-dimensional, pain-related experiences. Undoubtedly, additional measures are useful (and I tend to use the Gracely intensity and affective scales when possible) but the ease of administration of the VAS often trumps the time and effort of using longer tools (such as the McGill Pain Questionnaire) without much being lost.

Another troublesome aspect of the VAS is its proneness to ceiling and floor effects in studies with repeated measures. A participant rating a 9/10 at baseline cannot increase to a 12/10; a participant rating a 1/10 cannot decrease below 0/10. Therefore the relationship between the rating and the sensory experience is distorted at the edges of the VAS, which is a problem in analysis.

Using open-ended ratio-scales and anchoring stimuli (in experimental pain) might help to overcome this issue.

Given that one's sense of pain is subjective and that a pain perception (of a same-pain) experience may vary markedly between patients, the best a VAS can do is provide practitioners with a sense of suffering. A trio of VAS can serve both the patient and the practitioner quite well; Example- "In the past week and including this visit, give a best estimate of the Worst, Least and Present pain you have experienced."

If one uses a 10mm VAS scale the estimates can be measured, quantified and used as a means of better understanding a patient's response to treatment and assisting the practitioner in an immediate and/or future course of action.

Sufficient studies of the VAS for pain appear to have established a measure of reliability. ec

Ben Vancranburgh Interpretation

noci#s , nociception,pain perception and pain cognition

is a good Point for pain measuerment

In a word, "Yes." the VAS gives us numbers, curves and stats...but more importantly, they are...for the present, and perhaps the future, a best window into the unique experience and 'nephesh' of a pain-suffering soul. Ed

The simplicity of VAS and the NRS, as some of the commentators have highlighted, cannot be overemphasised. the 'subjective' nature of these tools of measurement in research setting, can be overcome in a properly designed double blind randomised trial to limit any potential bias and still able to achieve reasonable objectivity. I believe researcher will continue to use the tool until a cheaper alternative which those of us in poor resource community can harness.

VAS is not measuring pain , but pain intensity ,for one patient comparing to himself..In cancer pain we use it in daily routine... In my country those who are against the VAS , do not assess at all the patients, but relies to the feeling or clinical intuition...and that is bad for the cancer patients...they are not spontaneously complaining about pain, as pain have a special meaning , as worsening condition and so.Before discussing VAS , let's assess all the patient in routine!

Bravo Marilene. Thats exactly what I meant. In Italy there is a State Law that oblige caregivers to evaluate pain and to register the evaluation in the patients' chart. Shame we neede a law for that....

VAS, NRS and similar scales are simplistic approaches to access pain. A brief look at psychophysics of other sensory systems (e.g. visual or auditory) will reveal that they are too simplistic. However, they are robust and reliable (and with that I mean very conservative, very insensitive and unidimensional). One major concern is the reduction to a continuum of very few numbers (0-10 Likert scales), which do not allow any analysis of scale properties (Boaz has referred to that suggesting to use non-parametric startistics, but this is adding insensitive statistics to insensitive scaling, which will not improve the situation). The whole issue of pain measurement is a very simplistic one, but let us stay with the VAS scale or similar. We have shown previously (Magerl et al. Pain 1998, Rolke et al. Eur J Pain 2006, Rolke et al. Pain 2006, Magerl et al. Pain 2010) that one major problem lies in the obvious violation of the general law of psychophysics (Stevens power law), which necessitates the processing of pain ratings in log-space. This needs at least scales that encompass more than one order of magnitude (e.g. 0-100 NRS, or the fabulous Borg scales used in exercise research). Analysing distribution properties reveals log-normal distribution of pain rating data as well as pain thresholds. Unfortunately, this is not even considered in most pain studies (and cannot be tested in primitive 0-10 scaling !!). However, this would be in anyone's interest, since caring for distribution properties of data, then transforming to achieve secondary normal distribution will add greatly to sensitivity of comparisons, controls for outlier-related false estimates of mean etc. Not least, it leads to homogeneous distribution of variance (homoscedasticity), which is a precondition for proper analysis of variance. Another issue related to 0-10 scales is the implicit idea that we have to restrict it to natural numbers. This is indeed just inadequate, and any kind of partitioning to decimals is legitimate to improve sensitivity. Marilene mentions that VAS measures pain intensity, but this depends very much on wording of the top end of the scale (the word "worst", which is commonly used has a clear affective connotation). Commonly, pain affect is approached by asking for unpleasantness. But this is biased by the anglo-american angle of view (unpleasant is a much severer word in English than its translations in other languages, e.g. in German, Hebrew, or Asian languages, where unpleasant means something really minor). Thus, both dimensions correlate strongly. A much more relevant and upcoming concept will be the scaling of the suffering dimension (you may look up Smadar Bustan's work for that aspect). This will really express what is of relevance for a patient.

VAS is a scale only to be used to see/ measure outcome provided that the patient understand it. The NRS is better in this case for order people.

As an epidemiologist, I love the VAS. As a pain clinician, it is actually less useful. When I ask my patients, with a carefully worded question, to rate their pain intensity on a 0-10 scale, approximately half will give me a score >10!

And in one of our qualitative papers, our title is,"I feel so stupid...", this based on asking a number of older adults to score their pain, but being being unable to do so and feeling "stupid" because they couldn't meet the interviewers (apparently basic) expectations.

I meant NRS, not VAS - sorry!

Good!

But the older people in our study could still not all respond to an NRS. See attached.

Article “I feel so stupid because I can’t give a proper answer…” How...

So long as you set boundaries for 0 and 10 it is valid and nobody can offer

anything that is better

Therefore VAS lives and reigns

This problem is particularly evident in paople who cannot speak and use their reason. In these cases, VAS can be used by third parties, but the subject cannot give his/her assessment. Thus, an indirect evaluation of pain can be given by using other and comlicated pain scales, that use simultaneously several parameters. These scales have been criticised for their complexity and for the scarce need of scoring pain during acute pain. Pain scoring is useful during cronical pain, but in the case of acute pain such as injections, tracheal aspirations, heel-pricks and so on, it has little utility. The main problem is that all these scales decontextualize pain, and assess pain scores without any reference to the type of painful stimulation. I believe that in the case of provoking acute pain in babies, it is sufficient to be aware of the risk of provoking pain with reference to the type of stymulus and to the part of the body (and its state) where it is apllied. Then, we should see the reaction of the baby, but without any necessity of scoring pain. Any pain provoked to a baby is always a failure. 

The older person has difficulty in reducing and/ or defining their multiple sites and types of pain into a single numerical value which places an emphasis on a singular location of pain intensity. This is compounded by the fact that given the nature of pain not all types and sites of pain may be present within that period of time when the questioning is being done. Furthermore a reliance  a single numerical value as a holistic representation of the nature of the pain that is present within multiple types and sites of pain in older people does not provide individualised continuity for them.

Dear Joanne. It is a known fact that pain measurement at age extremes, is challenging  and difficult. For these we have other possible measures starting with the faces scales. Different types of pain must be evaluated with appropriate measures (e.g. DN4 for neuropathic pain). If you know that your patient have intermittent/remittent or oscillatory pain over time you should use pain diary. Finally, a thorough pain evaluation does not include only its intensity (measured by the VAS or NRS) which is only one dimension of the pain experience (i.e. one-dimensional measure); you should repeatedly evaluate, temporal, qualitative and topographical dimensions. Usually patients with several pain sites will report the most painful one on the VAS/NRS; treating that one will probably affect the others if they are of the same type, otherwise you have to tailor therapy for each pain site.

Thank you Boaz Samolsky Dekel, I agree totally 1000% that a comprehensive and through assessment of pain can never be underestimated.  I also thank you for your input relating to my previous post.

My concern for how an older person views a VAS scale or any other pain numerical value has arisen from my nursing PhD research on how nurses in acute care assess and manage pain in older hospitalised inpatients. 

In the interviews , the  older  inpatients spoke of their difficulty in expressing their pain not only into a singular numerical value, but they also talked about how they found it difficult when asked to compare types of pain and provide a numerical value for each.

VAS and other single item instruments in pain diagnostics and pain research are useful tools. But they often bear also severe problems.

Objectivity: Despite they are easy, situations of application and the verbal instructions of diagnosticians vary. Therefor objectivity represents a problem, which in clinical studies often is tried to be solved by using VAS as a part of a questionnaire.

Reliability: Reliability seems good for the intrapersonal use but for the comparison of persons its interpersonal comparability is unclear. This probably can be enhanced also by the use as a part of a questionnaire, with a defined instruction.

Validity: The main problem is validity. I agree with Stuart, that pain is largely a subjective phenomenon. And it contains all the problems we know out of the history of introspection methodology to measure it. On they other hand terminological questions are of central relevance. What is the difference between headache pain and a pinprick pain?

There are additional cultural problems, because pain concepts are partly cultural and subcultural bound, especially pain tolerance. Often there is more than one word for unpleasant experiences as well as terminological overlap of terms an concepts in different languages.

Another problem is the interpersonal situation in which the VAS is used. In one study we used the Inventory of Interpersonal Problems and found a significant correlation. We can conclude, that VAS measured pain has an interpersonal aspect. VAS is used by patients as an interpersonal message.

Dear Tony, indeed "patients are not stupid " and can remembre the spot on the VAS if they want.

Just a quick note on the distinction of pain intensity and unpleasantness. Intensity is probably clear, but unpleasantness is a term that is linked to Anglo-american language. ... and indeed unpleasantness means something serious in this tongue. Nevertheless, there is a very substantial correlation with unpleasantness. However, the distinction is still relevant as for e.g. in imaging correlation with unpleasantness defines separate brain areas, in psychophysical analysis they may differ, In many languages, however, e.g. German, Hebrew and most Asian languages,unpleasantness is not the same construct. First it is frequently not understood what that may mean (appart from intensity) up to the extent that patients may feel ridiculed if they are asked for a rating of unpleasantness of pain. What is urgently needed is a concept that reflects the severeniss of pain experience clinically, namely a valid estimate of a patient's suffering. And again, I doubt, that a third VAS for the dimension of suffering will do, but as a crude first attempt we should add this to our repertoire of global estimates.