What is fear? How is it linked to pain? What can it tell us about the outcomes we see in pain? Is it the big deal we have been lead to believe it is? Through this blog I posit that we don’t really know how fear is linked to pain, we haven’t exactly defined it, we can’t even measure it well much less use this model to predict anything much.
Science is theory driven and theories drive our understanding, but the process of science also affords us the luxury of throwing away these theories if they become redundant. The Fear Avoidance Model (FAM) wasn’t formulated in its modern guise until relatively recently amongst the scientific enquiry of pain. It was postulated by Lethem et al in the 1980s and was based on well established science around avoidance behaviour (Ref). It states, with assumptions that come along with the school of operant conditioning (that much of human behaviour is driven by the consequences of that behaviour);
“A threatening pain producing situation elicits a conditioned response of sympathetic activation including fear, which in turn leads to avoidance of the situation. Avoidance is then continually reinforced by a reduction in the painful or unpleasant stimuli leading to ‘disuse’ syndrome (Ref).”
It has since rollicked from strength to strength and has taken up almost mythical status in the modern musculoskeletal realm. For over 20 years, it has been proposed as the primary causal model for the development and persistence of disabling low back pain. Readily referred to by clinicians and policy makers; it was the great hope of pain science. Until it was examined by people who really knew how to look.
The link isn't clear
The association of pain to fear is described dating back to Aristotle; it’s not new, but it definitely isn’t simple. Pain related fear is different from catastrophisation and catastrophisation is very different to avoidance; these are all separate constructs amongst negative pain schema that are inter-related in complex ways (Ref). If fear leads to chronic pain, we may assume that individuals who have developed chronic pain would report high levels of fear; but that’s not what we see borne out empirically. The relationship is unclear and data shows us fear is not universally present in patients with chronic pain; Karoly and Ruehlman (2006) performed a cross-sectional analysis of ‘resilient’ and ‘non resilient’ chronic pain patients (n=526) and found a significant proportion of people reporting low levels of pain related fear (Ref). Similarly, with large prospective cohort studies, the reporting of pain related fear is evenly spread amongst groups and not clearly delineated amongst the proportion of people that do go on to develop chronic pain (Ref, Ref). This doesn’t mean that it isn’t a determinant for chronic pain development in some way or that we shouldn’t measure it, simply that we should place doubt where the FAM would have us place certainty.
The other constructs in the FAM don’t seem to be simply related either; noteably the relationship between fear and avoidance behaviours. This could be because human behaviour does not exist in a motivational vacuum. We avoid bodily harm or pain amongst a context of other, often competing goals (Ref). Indeed, when robustly studied, competing goals can change or stop avoidance behaviour (Ref). The FAM depicts clarity where there may be none and these links seem to fail to stand up to scrutiny when empirically analysed.
We can't even measure it
Central to this confusion could be because we have a hard time defining what we are actually talking about. The FAM intimately linked the concepts of fear and movement and we seem to have struggled to shake the confusing kinesiophobia concept ever since (Ref). It is also equated to other phobic disorders, which are represented by an irrational fear of typically neutral stimuli. Whereas pain is, in and of itself, a very threatening stimuli and fear of this can be viewed as completely rational (Ref). At odds with characterizing pain-related fear as phobia, questionnaires commonly associated with the FAM (e.g. Tampa Scale of Kinesiophobia (TSK), Fear Avoidance Beliefs Questionnaire (FABQ)) focus on pain-related beliefs and expectations while neglecting crucial attributes of phobic disorders like negative affect or distress-related behavior (Ref). The FABQ has come under scrutiny and in a cohort of 722 sick-listed people it seems to lack construct validity; unable to actually measure beliefs about physical activity (Ref). And other commonly used scales in FAM, like TSQ and the Pain Catastrophizing Scale (PCS) seem to lack responsiveness and are unable to discriminate between true change and random error (Ref). One has to ask, what is it that we’ve been targeting all of these years, if we can’t even define or measure it?
It can't predict outcomes
If we could measure it, the FAM creates an intuitive illusion that we could use it to predict outcomes. Once again though, when put under scrutiny, the FAM doesn’t seem to hold up to rigorous studies designed to ask predictive questions. There are tools designed to capture constructs of the FAM and categorise based off their measurement; one of these is the STaRT Back Screening Tool (SBST). Its predictive validity seems quite limited in low back pain in a recent prospective cohorts by Kendell et al (2017) (n=264). Another tool designed to capture FAM constructs and predict outcomes based on these is the Orebro Musculoskeletal Pain Screening Questionnaire (OMPQ). In a systematic review and meta-analysis of the predictive ability of this and STaRT back by Karran et al (2017) it seems we aren’t very good at predicting much at all when it comes to the development of chronic pain. Both outcomes measures (SBST and OMPQ) performing poorly in the ability to discriminate pain outcomes with pooled data of around 1153 and 360 respectively (Ref).
It doesn't explain outcomes
Much of what the FAM proposes isn’t actually concerned with prediction; instead it represents a causal relationship between fear, catastrophising and chronic disabling pain. This is fundamentally different to the ability to predict the onset of chronic pain. If two variables lie on the same causal pathway and one exposure causes a second exposure, which itself causes the outcome of interest, the second exposure is called a mediator (Ref). If the theory were valid, we would see elements of the FAM (that we can’t measure) like fear avoidance beliefs mediate the outcome in pain populations. This is not what we see. When studies that ask causal questions were pooled and analysed in a landmark systematic review by Lee et al (2015), it seems fear does not mediate the relationship between pain and disability. Further work has since demonstrated the poor potential that fear has to mediate (explain a causal pathway) in the outcomes of both disability and pain (Ref). Put simply, we don’t know how pain leads to disability and it’s unlikely that the FAM explains much of this pathway.
What do we do instead?
If not the FAM, then what explains this pathway and what can we do about it? There is something that has been demonstrated as more valid in both explaining the pathway to disabling pain and potentially something we should focus more of our attention to. This is self-efficacy, defined in the context of pain as,
"the beliefs held by people with chronic pain that they can carry out certain activities, even when experiencing pain” (Ref).
In Lee et al (2015)’s meta-analysis, self-efficacy outperformed elements of the FAM by some margin in explaining the development of disabling pain. This creates a common sense link between elements of interventions for pain and disability that we know create a larger effect than others - say exercise and graded exposure - and why that might be so. Once again though, this relationship may be closer to explaining why some people end up with disabling pain, the relationship is likely not simple and our interventions need to be complex to attend for this. Put another way, if exercise was the panacea then self-efficacy would explain the entire relationship between pain and disability; it’s not and it doesn’t.
With some fuzziness of the FAM dealt with, we can get onto proposing alternate theories (also to be debunked in time). Perhaps we need to continue to measure this, but do so with a healthy skepticism about the conclusion we draw. Potentially fear of pain is simply a rational reaction and instead of tackling this head on (which could lead to perceived invalidation), we can accept and understand this fact both as practitioners and patients. Potentially we should focus more on fostering the belief of an ability to carry out meaningful activities rather than put misplaced energy into reducing an unpredictable and irrational emotion (fear). Perhaps if we focused on these things, we may be better at preventing disabling pain.
Although there has been recent valiant attempts to make sense of all of this confusion regarding the FAM and it’s value in the clinical progression of low back pain (Ref). And although the FAM may be useful in some regard, the current data demonstrate it isn’t useful in the sense we are using it; great hope no more. Although fear is evidently linked to pain, when scrutinised the FAM seems to be unable to account for the multitude of non linear, complex relationships that pain and disability come along with. The scientific process affords us to update our beliefs when necessary and it’s probably time we update our beliefs in this regard. It’s time we update this model; we’ve got a long way to go until it will tell us anything useful about the outcomes we see.