What’s the biggest barrier to learning more?

Reading and engaging with clinicians online and face-to-face, it’s clear to me that effectively integrating psychosocial factors into daily clinical reasoning, especially amongst physical or manual therapists, is a real challenge. There’s enough research around showing how poorly these factors are identified and then factored in to change what we do and how we do it for me to be convinced of this. What intrigues me, though, is why – given psychosocial risk factors have, in NZ, been around since 1997 – it’s still a problem.

It’s not ignorance. It’s not holding an alternative viewpoint. It’s not just that clinical reasoning models don’t seem to integrate these factors, or that our original training kinda partitioned the various “bits” of being human off – I think that it’s probably that we think we’re already doing well enough.

Image result for dunning kruger effect

This effect has a name – Dunning-Kruger effect. Now, don’t be put off by this term, because I know in some social media circles it’s used to bash people who are  maybe naive, or haven’t realised their lack of knowledge, and it can feel really awful to be told “well actually you’re ignorant”, or “you’re inflating your skill level”.  The thing is, it’s a common experience – we all probably think we’re great car drivers – but in reality we’re all pretty average.

The same thing occurs when we consider our ability to be:

  • empathetic
  • responsive
  • good listeners
  • client-centred
  • collaborative

Another important effect found in clinicians is that we believe our experience as clinicians means we’re better at aspects of clinical care, and especially at clinical reasoning. Over time we get better at recognising patterns – but this can actually be a problem for us. Humans are excellent at detecting patterns but as a result we can jump to conclusions, have trouble stopping ourselves from fixating on the first conclusion we draw, begin looking for things to confirm our hunch, overlook things that don’t fit with the pattern we’ve identified, and basically we begin to use stereotypes rather than really looking at the unique person sitting in front of us (see Croskerry, Singhal & Mamede, 2013a, b).

The effect of these biases, and especially our bias towards thinking we do better than we actually do (especially regarding communication skills and psychosocial factors) means we’re often completely unaware of HOW we communicate, and HOW poorly we pick up on psychosocial factors.

So often I’ve heard people say “Oh I use intuition, I just pick up on these psychosocial issues” – but the problem is that (a) we’re likely to over-estimate how well we pick up on them and (b) our intuition is poor. The risk for our patients is that we don’t identify something important, or alternatively, that we label something as a psychosocial risk factor when it’s actually irrelevant to this person’s problem.

Clinical reasoning is difficult. While recognising patterns becomes easier over time because we have a far broader range of patterns we’ve seen before, at the same time

  • research is expanding all the time (we can be out of date)
  • we can get stuck prematurely identifying something that isn’t relevant
  • we get hooked in on things we’ve just read about, things that happen rarely, things that remind us of something or someone else

Hypothetico-deductive reasoning is an alternative approach to clinical reasoning. It’s an approach that suggests we hold some ideas about what’s going on in our mind while collecting more information to test whether this is the case. The problem here is that we look for information to confirm what we think is happening – rather than looking for something to disconfirm, or test, the hypothesis we hold. So, for example, we might observe someone’s pain behaviour and think to ourselves “oh that person is doing that movement because of a ‘dysfunctional movement pattern’. We can assume that the reason for this movement pattern is because of underlying dysfunction of some sort – but we fail to test that assumption out to see whether it might in fact be a movement pattern developed because someone told the person “this is the way you should move”, or the person is moving that way because of their beliefs about what might happen if they move differently.

The problem with intuition and these other cognitive biases is that they simplify our clinical reasoning, and they reduce effort, so they’re easy traps to fall into. What seems to help is slowing down. Deliberately putting a delay in between collecting information and making a decision. Holding off before deciding what to do. Concurrently, we probably need to rely less on finding “confirming” information – and FAR more on collecting information across a range of domains, some of which we may not think are relevant.

That’s the tough bit. What we think is relevant helps us narrow down our thinking – great for reducing the amount of information we need to collect, but not so great for testing whether we’ve arrived at a reasonable conclusion. My suggested alternative is to systematically collect information across all the relevant domains of knowledge (based on what’s been found in our research), wait a bit and let it settle – then and only then begin to put those bits and pieces together.

Why doesn’t it happen? Well, we over-estimate how well we do this assessment process. We do jump to conclusions and sometimes we’re right – but we wouldn’t know whether we were right or not because we don’t check out alternative explanations. We’re pushed by expectations from funders – and our clients – to “set goals” or “do something” at the very first assessment. We feel guilty if we don’t give our clients something to take away after our initial assessment. We want to look effective and efficient.

Great quote?

For every problem, there is a solution that is simple, elegant, and wrong. H.L. Mencken.

If you’d like to question your own practice, try this: Record your session – and transcribe that recording. Notice every time you jump in to give advice before you’ve really heard your client. Notice how quickly you form an impression. Examine how often you look for disconfirmation rather than confirmation. See how often you ask about, and explore, those psychosocial factors. It’s tough to do – and sobering – but oh how much you’ll learn.

Croskerry, P., Singhal, G., & Mamede, S. (2013). Cognitive debiasing 1: origins of bias and theory of debiasing. BMJ Quality & Safety, 22(Suppl 2), ii58-ii64. doi:10.1136/bmjqs-2012-001712

Croskerry, P., Singhal, G., & Mamede, S. (2013). Cognitive debiasing 2: impediments to and strategies for change. BMJ Quality & Safety, 22(Suppl 2), ii65-ii72. doi:10.1136/bmjqs-2012-001713


  1. Bronnie, I suggest that these observations extracted from a blog* (a conversation with Professor Horst Ruthrof) identify the huge problem that we face as clinicians:

    “Science and normativity

    Beginning with a broad perspective, a distinction that appears to me to be relevant to the characterization of the two domains of the clinician, on the one hand, and of persistent pain sufferers, on the other, is that between a scientific focus and one informed by normativity.

    The former can be classified traditionally as committed to the data of empirically observable evidence systematized via the testing of hypotheses and procedures of deduction and induction.

    This is the essence of the scientific method upon which modern medicine is so reliant.

    Normativity, in contrast, can be regarded as a summary bag containing all the sort of things that science, as science, is not designed to deal with, such as ethical, aesthetic, religious, and political values, as well as ideological dispositions. While in social practice the two domains always overlap, the principles that inform each are plainly different. It seems to me that it is the clash of those principles that make medicine an ethically highly complex arena of human endeavour.

    This clash also explains why so many sufferers of persistent pain are perplexed and disappointed when their clinicians appear to be deaf to some of these important issues in their lives.

    The social expectations about clinicians appear to be that they are trained primarily to apply scientific principles to the injured human body in order to return it to functional health. Persistent pain throws a big spanner in this social mechanism.”


    1. Some great points here. Within our communities there is an expectation that people will fit into normative ideals. We’ve always had “deviants” or people who don’t meet those expectations, and various communities have found ways to deal with them – from honouring them as prophets or soothsayers or oracles, to denigrating them and shunning them as mad or bad.

      People living with pain can often violate those normative ideals – yet at the same time, characteristics of people within this persistent pain community also fall into norms. There is the person who diligently works away at everything he or she is told. The one who constantly seeks reassurance. The one who looks for a cure. The one who gives up. The one who accepts and works within his or her limitations.

      At least part of the problem is, I think, the complexity of who we as health professionals are serving. Do we serve the person who just wants the pain to go away? And what do we do if we can’t make that happen? Do we serve the insurer who wants the claim to be closed? And what do we, as health professionals, do with those who can’t or won’t meet the criteria for claims closure? What level of responsibility do we have towards the community vs the individual? The individual vs her family? Do we respond to the pain and distress, or the disability, alleviating one but prolonging the other?

      I think it’s possible to incorporate normative values AND honour the individual, while also respecting the part we play within a social network. We can draw on scientific understanding – as far as it goes right now (accepting that we only approximate knowing what is “real” and what is not, and hoping we will get closer over time) – we can also honour the unique values and experiences of the person we work with, using the scientific knowledge to help the person and ourselves begin to create a narrative framework within which change is possible and measurable, and in doing so, help the person develop their own understanding of the contributors to their experience. We can recognise that this interaction exists within our current social context – and be prepared to shift as social values shift.

      People living with pain AND clinicians AND society ALL expect that treatment “should” be available to eliminate pain, reduce suffering and reduce disability. We all bring these expectations to our clinical consultations. Perhaps this is something researchers can highlight as a problem for our communities, patients and ourselves. The next step is developing ways to be OK with our own limitations (all of us – clinicians, patients and the community). We do this in end of life care – but we don’t tend to do this for people living with ongoing disabling illness, particularly pain.

      HOW do we change these expectations? And what is the fine line we walk between doing nothing (because we know we’ll be unsuccessful), doing too much (because we hope we’ll achieve it for some people, maybe this one), and sitting somewhere in the middle where we foster hope of helping people express who they are despite their limitations? In this I think we have a duty of care as clinicians to
      (1) let people know we’re there to help them express important values and activities (sense of self) despite their pain while still working to influence those things we know from research could have an impact
      (2) let funders/insurers know that for some people, “return to work” is not an achievable or even an admirable outcome, and that pursuing this is unhelpful and wastes resources,
      (3) work within our professions to create maturity and responsiveness to our limitations and finally
      (4) promote community understanding that being flexible does not equal being unfair, it simply means people hold different values about what is important.


    2. Sounds about right! Which I think was David Hume’s critique also! “claims about what ought to be on the basis of statements about what is. Hume found that there seems to be a significant difference between positive statements (about what is) and prescriptive or normative statements (about what ought to be), and that it is not obvious how one can coherently move from descriptive statements to prescriptive ones. The is–ought problem is also known as Hume’s law, or Hume’s guillotine.” https://en.wikipedia.org/wiki/Is%E2%80%93ought_problem

  2. I think that changing the way we do things really requires a paradigm shift to be comfortable asking psychosocial information. It is within our domain as it is part of the patient we are seeing. i feel many clinicians do not have outcome measures because they are too busy seeing too many patients. So we need to have a shift there as well. from what I see often the new clinician wants to see more patients as they make less money per patient and often see the patients on government insurance which also pay less. as clinicians we need to speak out as advocates for ourselves and give the patient ample time to get to the these issues. when I am teaching students this is often their worry about psychosocial issues and the outcome measures. they are not given the time and they feel they should conform as new grads so they are getting the info but not feeling comfortable asking for enough time.

  3. Even if we focus on “collecting information across a range of domains”, wont we just end up with more noise, (so-to-speak)? The problem with the BPS is that it doesn’t tell us how to prioritise, … “guide, recommend or restrict which features should be evaluated in any domain. The clinician is free to choose from a variety of potential tests, so approaches to patient evaluation risk reflecting the professional or attitudinal bias of the clinician. Neither does the model inform on how one domain may or may not influence or interact with another domain.”

    1. “The model” is, as John Quintner often reminds me, is not a “model” in so much as a theoretical model providing testable hypotheses. BUT by using this framework, together with the person and knowing the research, allows us to collect information across multiple domains – and use the mid-level theories to build a set of hypotheses about what might be occurring in this person’s experience at this time. So, for example, we do know that social responses often provide positive reinforcement for behaviours – so we can test this in the context of the person we’re working with. They may notice that family members go from solicitous to punitive – and reflect on their response to these behaviours, and can then decide for themselves whether this is working for them, or not. Or in what contexts these things are working, or not.
      We could draw from Vlaeyen’s “pain-related fear and avoidance” model to see how/whether this is influencing the person… and combine this with our understanding of nervous system plasticity to generate some hypotheses that we can then test to see how well they predict what’s happening for the person.
      My approach is to review research, and put it to the person as some possibilities – and together we collaborate to see how well these explain their situation.
      I do approach pain and pain management from a broad perspective, across the multiple domains we suspect may be involved. Irrespective of how “accurate” any model of experience is, I think we will always need to accommodate and adjust for the unique person we see in front of us. So, I hold the “model” lightly as a set of possibilities rather than probabilities. I work with the person as much as possible. I use the broad knowledge base I have to generate multiple possible explanations and collaborate with the person who is the ultimate judge of whether something is helpful or not.
      I’ve found that the broad domains we’re learning from the last 40 years of research into pain are really helpful, but none of them completely or fully explain why this person is presenting in this way at this time – that’s why I suppose a BPS model is a framework, and hypotheses are tentative.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.