Bias in Clinical Experience: Good intentions pave the road to mistakes

Hello again. In this post I intend to continue my exploration of why we need to understand the scientific method, why we need to inform our practice in research papers and how we can balance this with our personal/clinical experience in daily practice in a way that is beneficial to our patients.

If you haven’t already, please check my two-part introduction to evidence-based practice, starting here and then here. In these two texts, I have briefly covered the history and definition of evidence-based practice, what elements constitute it and how the funnel model helps us apply evidence to form useful knowledge.

In the first part, I also argued why what we experience with our own two eyes and ears during daily practice, although important, may not be enough to ensure we are doing the right thing to help the people who put their health in our care. A key element in this was the concept of bias, which is exactly the topic I want to focus on in this text.

I just want to make a brief disclaimer that, despite what may be frequently perceived, evidence-based practice does not dismiss clinical experience. It actually aims to better integrate research evidence, clinical experience, and expertise, as well as patient’s preferences and circumstances (Finch, 2007; Kamper, 2018(a); Meira, 2020; Sacket et al, 1996). With this aim, evidence-based practice challenges clinicians to accept there may be flaws to their reasoning and identify their own biases (Kamper, 2018(a); Sim and Wright, 2002).

To show that this is something quite normal, I’ll start with some of my own biases:

• Strength training is one of the most important forms of exercise we should all do.
• Coffee is healthy and good for us.
• Video games are not a bad hobby.
• In clinical practice, listening to and understanding patients is far more important than we do to them.

I’ve checked and after this, I still have all my limbs attached, I haven’t lost my job, the world hasn’t ended and more importantly, all my books and video games haven’t spontaneously combusted. That’s reassuring. But it shows that this humbling process isn’t harmful, even when I identify that some of the things I’ve listed may end up being wrong or that some of you may disagree with them.

But let’s ask: why is bias relevant to us?

As I explored in the previous posts, biases are a form of problem-solving process that our minds use, called Heuristics (Monteiro et al, 2020; Richarson, 2014; Saposnik et al, 2016). They are mental shortcuts that make our decision-making a bit simpler and more energy efficient. However, due to being quicker, we end up not spending as much time analysing information, thus making more mistakes (Phua, Fams, and Tan, 2013).

 If we ask an epidemiologist what is the definition of bias, they would say something like “a systematic deviation of the sample parameter estimate from the population value.” (Kamper, 2018(b)). Even though we’re all nerds here (c’mon, you’re spending your free time reading a text about evidence and critical thinking. Don’t deny it, you’re awesome), that’s a lot of big words.

Using more common language and applying it to the context of clinical observation, bias can be defined as “a difference between the results we see and what is actually happening” (Kamper, 2018(b); Sim and Wright, 2002).

To better help us understand, let’s look at some of the biases that are more common to occur when we try to interpret our personal and clinical experiences from day-to-day practice.

Confirmation bias

Imagine a physiotherapist who believes they are good at their job and that they have their patient’s best interests always in consideration. If not the first part, most of us will likely see ourselves reflected in the second part of that statement. And there is nothing wrong with that.

However, after further studying the human man, we have learned that we tend to, unconsciously, pay more attention to information that supports our beliefs or preconceptions, while simultaneously both interpreting ambiguous information as confirming our beliefs and ignoring information that refutes or contradicts said beliefs (Kamper, 2018(a); Monteiro et al, 2020).

This being the case, how do you think this will affect the above physiotherapist’s remembering of how their treatments affected their patients?

Evaluation apprehension bias

When someone is being tested, such as during any kind of clinical assessment, they may feel anxious to a certain degree. This can lead to them to try to answer or behave as they believe is expected by the assessor, instead of how they would normally (Bowling, 2014).

How do you think this will affect the results of the assessment?

Interviewer bias

When assessing or interviewing someone, we can unconsciously ask leading questions or express ourselves to be a certain type of clinician or person. This in turn will influence how someone will answer our questions during assessment or subsequent follow-ups. (Bowling, 2014)

Observer bias

Generally, due to our perception of things and way of thinking being affected by our beliefs and emotions, as well as some degree of variation every time we do the same thing, there are differences in how something really happened and how we perceived it as happening (Bowling, 2014; Phua, Fams and Tan, 2013).

Recall or Availability bias

We all have the tendency to quickly forget the more common events we see but remember very well the out-of-the-ordinary events, such as when something goes really well, or really badly (Bowling, 2014; Kamper, 2018(a)).

In our clinical practice, this leads to us basing our future decisions on those patients who did very well or very badly, instead of those with a more average, but also more frequent, outcome (Kamper, 2018(a)).

Selection bias

When working with our patients, the characteristics of those patients will be affected by things such as our geographical area, not only because of the geographical terrain, but also the socio-economic and political context of that area. Because of this, the characteristics of our own patients may differ from that of the wider population with that same condition or care needs. Because of how selective our patient group is as a result of this, we can’t really generalise what we observed with them to the rest of that population (Kamper, 2018(a)).

In addition to these specific documented biases, there are other issues we may come across that lead us to make errors in the interpretation of our clinical observations. When a patient improves following treatment it may be due to the treatment itself, but this improvement may also be caused by other factors that have an effect at the same time that treatment is performed (Kamper, 2018(a)). Let’s explore some of those other factors.

Natural History

Often health problems, if they’re not serious conditions, will resolve on their own with just the passing of time, independently of whether someone receives treatment or not (Kamper, 2018(a)). This can lead to us treating someone and seeing them getting better because our intervention coincided with the timing of recovery for that condition through its natural history.

Regression to the mean

It’s common for health problems to fluctuate in how they present and feel. This means that often our patients can feel like they have periods of time where they have very mild or no symptoms at all, with episodes of very severe flare-ups in the middle (Kamper, 2018(a)). People with this type of condition tend to book appointments with a healthcare professional exactly on those episodes when their pain becomes worse. Because of this, the next fluctuation they will experience is that improvement towards more mild symptoms, appearing that they improved after their appointment because of what was done in that appointment (Kamper, 2018(a)).

Placebo Effect

A lot of people have probably heard of this effect by now. It basically describes how when we believe in something, such as that a treatment provided by a healthcare professional will make us better, that expectation will have a physiological response in line with that belief – we will feel better (Kamper, 2018(a)). However, it is important to identify that even though placebo effects are often linked to interventions or treatments, they are not part of the effect of said interventions or treatments (Kamper, 2018(a)).

Polite patients

Most, if not all, healthcare professionals will try to build a positive relationship with people who comes to them for care (Kamper, 2018(a)). It’s easier to try to help someone if they like us and are willing to collaborate. But this can backfire in a way: if our patients really like us, they may want to avoid making us feel bad or useless and will tell us a little white lie of how they are improving a lot, or look and act as if they are much better, even if that is not the case (Kamper, 2018(a)).

On that anxiety and self-doubt-inducing note, I want once again to remind you that none of this means that our personal observations and clinical experience are useless. They are valid sources of knowledge and can help us make decisions about someone’s care. However, we have to recognise and understand how these sources of knowledge can be very faulty and lead us to commit errors.

I hope this text has made you consider keeping open the possibility that what you observe and experience may not be an accurate reflection of reality and that when you are presented with a robust and trustworthy piece of evidence that challenges what you thought was true, even if you don’t accept it straight away, at least consider in your mind: “What if I am wrong here?”.

I can assure you that despite how scary that may sound, a lot more positive than negative will come out of that consideration. It’s normal and healthy to make mistakes and change our minds. It’s not just our observations and thought processes that are flawed. Any piece of information, independently of coming from media, clinical experience, professional courses, colleagues, or even research, can be biased and have flaws (Kamper, 2018(a)). And it’s exactly the bias of that last one that I will be exploring in the next post.

Thank you for reading this far and I will hopefully see you nerds in the next one,

The Physiolosopher.

References:

Bowling, A. Research Methods in Health – Investigating health in health services. 2014, fourth edition. Open University Press. Berkshire, England

Finch, P. M. (2007). The evidence funnel: Highlighting the importance of research literacy in the delivery of evidence informed complementary health care. Journal of Bodywork and Movement Therapies, 11(1), 78–81. https://doi.org/10.1016/j.jbmt.2006.09.001 

Kamper, S. J. (2018 (a)). Engaging with research: Linking evidence with practice. Journal of Orthopaedic and Sports Physical Therapy, 48(6), 512–513. https://doi.org/10.2519/jospt.2018.0701

Kamper, S. J. (2018(b)). Bias: Linking evidence with practice. Journal of Orthopaedic and Sports Physical Therapy, 48(8), 667–668. https://doi.org/10.2519/jospt.2018.0703

Meira, E. (2020). Understanding evidence-based medicine using a funnel analogy. Society for Transparency, Openness and Replication in Kinesiology. https://doi.org/10.31236/osf.io/kr6aq 

Monteiro, S., Sherbino, J., Sibbald, M., & Norman, G. (2020). Critical thinking, biases and dual processing: The enduring myth of generalisable skills. Medical Education, 54(1), 66–73. https://doi.org/10.1111/medu.13872 

Phua, D. H., Fams, E., & Tan, N. C. (2013). Cognitive Aspect of Diagnostic Errors (Vol. 42, Issue 1). 

Richardson, L. G. (2014). Awareness of Heuristics in Clinical Decision Making. Clinical Scholars Review, 7(1), 16–23. https://doi.org/10.1891/1939-2095.7.1.16 

Sacket, D. L., Rosenberd, W. M., Gray, J. A. M., Haynes, R. B., & Richardson, W. S. (1996). Evidence based medicine: what it is and what it isn’t. BMJ, 312, 71–72. Saposnik, G., Redelmeier, D., Ruff, C. C., & Tobler, P. N. (2016). Cognitive biases associated with medical decisions: a systematic review. BMC Medical Informatics and Decision Making, 16(1), 1–14. https://doi.org/10.1186/s12911-016-0377-1