Explaining symptoms in general practice: a predictive processing framework for primary care consultations 2026 van Boven et al

Andy

Andy

Senior Member (Voting rights)

Abstract​

Aims​

To present predictive processing as a clinically useful framework for explaining symptom variability in general practice consultations.

Methods​

Development of a conceptual framework through an iterative process of theoretical synthesis and clinical translation, informed by a supporting literature search integrating literature on predictive processing and symptom perception with primary care literature on consultation practice and persistent physical symptoms.

Results​

Predictive processing conceptualises symptom perception as an interaction between bodily signals and the brain’s prior expectations. Symptoms remain real and potentially disabling even when pathology is stable or minimal. The framework provides a structured approach for consultations: validating patient experiences, explaining symptom persistence, and guiding recovery-oriented management. A worked consultation example illustrates clinical application.

Conclusions​

Predictive processing offers a coherent framework to help general practitioners explain symptoms in primary care. Its limitations, including restricted prescriptive guidance and limited direct trial evidence, are acknowledged. The framework supports patient understanding and guides management focused on gradual recovery.

KEY MESSAGES​

  • Patients with persistent physical symptoms in the absence of clear pathology frequently feel misunderstood; general practitioners lack a practical explanatory model that both validates and explains.
  • Predictive processing offers a generalisable mechanism through which symptom experience can be explained without medicalising or psychologising.
  • Clear explanation in the consultation reduces patient uncertainty, supports adaptive coping, and guides graduated recovery without unnecessary investigation.

WHAT THIS PAPER ADDS​

  • Integrates predictive processing theory with the biopsychosocial approach, specifically applied to consultation practice in general practice.
  • Provides a worked consultation example (ankle pain) illustrating practical application of the framework.
  • Identifies the limitations and evidence gaps of predictive processing as a clinical framework, and outlines priorities for future research in primary care.

Open access
 

A further concept is ‘precision weighting’​

The brain continuously judges how much trust to place in incoming bodily signals versus its own prior expectations. This process of weighing internal bodily signals against expectation is what is meant by interoceptive inference. In states of heightened anxiety or uncertainty, this balance shifts so that the brain relies more heavily on its expectations and becomes more sensitive to ambiguous signals, making them more likely to be interpreted as significant and threatening. The effect is most pronounced when peripheral signals are weak, ambiguous, or fluctuating, as often occurs in post-infectious states and functional syndromes, the brain relies more heavily on priors, and symptom intensity may be driven more by expectation than by detectable peripheral change. This pattern is recognisable across presentations common in primary care: patients with post-COVID fatigue whose exhaustion persists beyond resolution of the acute infection [Citation20], patients with post-infectious syndromes more broadly [Citation22], patients with functional dyspepsia whose symptom severity fluctuates independently of endoscopic findings [Citation23], and patients with chronic low back pain whose disability correlates poorly with radiological findings [Citation24]. These examples share a common mechanism: persistent or non-specific input from the body, whether residual physiological change, stress responses, or poorly understood sensations, interacts with the brain’s predictive model and strong prior expectations to shape what is perceived. Neither peripheral signals nor priors alone determine the outcome; it is their interaction, weighted by the brain’s assessment of reliability, that drives symptom experience. And it is precisely this interaction that the consultation can address.
Click to expand...
 
Box 4. Extending the framework beyond pain: persistent fatigue

Case

A 44-year-old man presents with persistent fatigue six months after a confirmed viral infection that has otherwise fully resolved. Blood tests and further work-up are unremarkable. He struggles to link his fatigue to any ongoing physical cause and fears his GP will attribute it to ‘being stressed’.

Explanation

  • “Your tests are reassuring, there is no ongoing infection or other disease process driving this. What is more likely is that your body, quite reasonably, learned to protect itself while you were acutely unwell, and your brain is still generating fatigue signals as a precaution, even though the original threat has passed.”
  • “This is not a sign that something has been missed, and it is not ‘just in your head’—it is a normal, if unhelpful, side effect of how the brain protects the body. You can help it update that expectation through paced activity: a graded, symptom-informed increase in what you do, rather than resting completely or pushing through until you crash.”

Commentary

This example illustrates that the same predictive mechanism described for pain applies to fatigue, where there is no visible ongoing injury to point to—and shows paced activity applied outside a musculoskeletal context.
Click to expand...
 
The senior author is from Belgium the other authors are from the Netherlands:

Kees van Boven
Department of Primary and Community Care Radboudumc, Nijmegen, Gelderland, The Netherlands,

Peter Lucassen
Department of Primary and Community Care Radboudumc, Nijmegen, Gelderland, The Netherlands,

Tim Olde Hartman
Department of Primary and Community Care Radboudumc, Nijmegen, Gelderland, The Netherlands
&
Omer Van den Bergh
Health Psychology, University of Leuven, KU Leuven, Belgium
Contribution
Supervision, Writing – review & editing
 
bafkreidlgllgoxojxlnaljpt3jog5c4dzoahq5tu7bgcsfjaodtmz5n2om.webp
Artist: Eleanor Fielding

The cartoon series is called - Out of Patients

There are three boxes in which a male doctor and woman patient are talking.

Box One - the doctor is saying “I'm afraid your illness is psychosomatic women’s stuff-you need exercise and psychiatry”
The woman is looking down at the test results and saying “I can't believe it..”

In Box two the woman is laughing and the doctor has crossed his arms in annoyance as she says: “..you're really offering medicine with all the science taken out!”

In Box three the doctor is bright red and the woman has turned toward the viewer to say “Quackery with Credentials ..available free at your local GP”

The cartoon maker is #FanningTheFl
 
From the limitations:
The framework also presupposes careful exclusion of serious disease but cannot eliminate the residual risk of missed diagnosis: once a predictive processing explanation has been adopted, new or evolving symptoms might receive less attention. Robust safety-netting and periodic review therefore remain essential, and the framework should be understood as an adjunct to, not a substitute for, thorough diagnostic reasoning.
Click to expand...
It’s fascinating that the authors (actually the AI) do not spend any time reflecting about the ethics of telling someone that nothing is wrong with them and that their beliefs and behaviours are the reasons for their symptoms, and encouraging them to not think of symptoms as a reason to go see a doctor, while at the same time knowing that something might actually be wrong with them.
It remains uncertain how individual patients respond to brain-based symptom explanations; some may still experience these as psychologising, despite the non-stigmatising framing.
Click to expand...
Apparently the bully is allowed to decide what is bullying!
The selection of literature was not exhaustive and inevitably reflects the authors’ interpretive judgements.
Click to expand...
They even admit that they have not tried to look at all of the evidence and to make things make sense across the board. This is just the interpretation they prefer.

Not only that, they got the AI to write it for them:
The authors used an AI-based writing assistant (Claude, Anthropic) during the preparation of this manuscript to support drafting, structural editing, and language refinement. All content was critically reviewed, revised, and approved by the authors, who take full responsibility for the integrity and originality of the submitted work. The AI tool is not listed as an author.
Click to expand...
Pangram classified 83 % of the full text as AI generated.
IMG_0146.webp
The only sections that are not written by AI are the two last bullet points in the limitations and the conclusion.
 
The selection of literature was not exhaustive and inevitably reflects the authors’ interpretive judgements.
Click to expand...
Is that what they're calling bullshit opinions nowadays?

It takes a whole system to fail for something like this to be commonplace, at this point for decades. It's a complete bullshit factory, whenever any bit of bullshit starts to waver, they just trot out some variation of combination of it. If not for the explicit context that gives everyone involves immunity, this would qualify as a criminal enterprise. It's an immiseration machine that produces suffering and, somehow, fulfilling academic careers.

Edit: In the methods section, they write:
informed by a supporting literature search integrating literature on predictive processing and symptom perception with primary care literature on consultation practice and persistent physical symptoms
Click to expand...
So with the first sentence I quoted, in which they admit they did not do an exhaustive search, they still claim that the (non-exhaustive) literature search informs it. Basically they're admitting that they went out to seek confirming citations. Incredible.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom