CRP Testing Drove Very Different Prescribing Behavior in UK and Norwegian Primary Care
C-reactive protein has long been promoted as a practical aid for antibiotic stewardship, but its real influence depends on how clinicians use it at the bedside. In a behavioural experiment spanning UK and Norwegian general practice, physicians responded very differently to the same clinical vignettes once CRP became available as optional information. The clearest signal came before any laboratory number even appeared: Norwegian GPs asked for CRP far more often than their UK counterparts.
The study tested a simple but important question. If a biomarker is meant to reduce unnecessary antibiotics, does it actually sharpen judgment, or does it become another cue that nudges clinicians toward a familiar decision pattern? In primary care, where respiratory and other infectious presentations are common and uncertainty is part of everyday practice, that distinction shapes both prescribing and antimicrobial resistance.
How the Experiment Simulated Primary Care Uncertainty
One hundred and one GPs in each country completed four patient scenarios covering a range of infectious or inflammatory conditions. After reading each vignette, they could request additional information, including CRP. The CRP result was randomly assigned as high or low for each GP and scenario, a design that allowed the investigators to observe how the biomarker itself influenced judgment rather than how it correlated with a patient’s underlying illness.
Each time new information was revealed, clinicians rated whether antibiotics were needed on a Likert scale. At the end of the scenario, they selected a final management option. That structure matters because it captures the stepwise way real consultations unfold: the initial impression, the selective search for more data, and the eventual prescription decision. Biomarkers rarely operate in isolation. They enter a cognitive process already shaped by symptom severity, diagnostic uncertainty, prior experience, and local prescribing norms.
The experiment therefore sits squarely in the behavioral space of antimicrobial stewardship. It does not measure patient outcomes, but it does reveal how a commonly used point-of-care test may alter clinical reasoning before a prescription is written.
Norwegian GPs Used CRP Far More Often Than UK GPs
Norwegian GPs requested CRP in 77% of opportunities, compared with 28% among UK GPs. The odds ratio was 18.09, with a 95% confidence interval of 9.86 to 33.20, and the comparison reached statistical significance with P less than the reported threshold in the abstract excerpt.
That gap is striking. It suggests that CRP was not treated as a marginal backup test in Norway; it was embedded in the clinical workflow to a far greater extent than in UK practice. In day-to-day terms, a biomarker that is promoted as a decision aid can become part of the default diagnostic reflex when a health system trains clinicians to rely on it.
The practical consequence is immediate. Once testing becomes routine, the question is no longer whether CRP exists, but how often it is sought before antibiotics are considered. In systems where clinicians are already accustomed to ordering it, the test may shape thinking earlier in the consultation and may influence the threshold for labeling illness as bacterial, inflammatory, or safe for observation.
Why a Biomarker Can Change Judgment Without Solving Diagnostic Uncertainty
CRP is biologically appealing because it reflects systemic inflammation, but inflammation is not synonymous with bacterial infection. That limitation is central to interpreting the trial. A high CRP result may intensify concern and push a clinician toward antibiotics, while a low result may reassure and support watchful waiting. Either way, the biomarker does not identify the cause of inflammation with perfect precision.
In primary care, that ambiguity can be useful or misleading depending on context. A test that reduces uncertainty for one clinician may simply reframe it for another. If the result is treated as an objective arbiter rather than one input among several, prescribing behavior can drift toward overconfidence. If, instead, the number is interpreted alongside examination findings and illness trajectory, CRP may support more selective treatment.
The design of this experiment was well suited to expose those cognitive effects. By allowing clinicians to choose when to request CRP and then randomizing the result, the investigators could see both the appetite for testing and the extent to which the result itself steered treatment thinking. The abstract excerpt does not provide the downstream antibiotic selection data, so the most defensible conclusion is narrower: access to CRP clearly influenced information-seeking behavior, and that influence differed sharply between countries.
What the Country Gap Means for Stewardship Policy
A point-of-care test does not work in a vacuum. Its impact depends on training, local practice style, patient expectations, and the clinical culture surrounding uncertainty. The Norwegian pattern suggests a system in which CRP has already been normalized as part of routine assessment. The UK pattern suggests a more restrained approach, or at least less habitual use.
For antimicrobial stewardship, that difference cuts both ways. Wider CRP adoption could potentially reduce unnecessary prescribing if low results consistently reassure clinicians away from antibiotics. It could also increase testing without meaningfully changing prescribing if clinicians order CRP simply to justify decisions they were already inclined to make. The behavioral experiment cannot answer that downstream outcome question on its own, but it does show that country-specific practice norms strongly shape how the test enters the consultation.
The numbers also raise a policy issue that is easy to overlook. Implementation is not the same as evidence. Even a well-established test may be used very differently across health systems, and those differences can alter its value as a stewardship tool. A biomarker promoted as a standardized solution may, in practice, amplify local variation unless clinicians are taught when the result should matter and when it should not.
Limits of the Available Abstract Data
The abstract excerpt provides a clear signal on test requesting but does not include the full antibiotic-prescribing outcomes, scenario-by-scenario results, or any patient-level clinical endpoints. It also reflects experimental judgment in response to vignettes, not real-world consultation behavior or adverse events. Those boundaries matter because vignette studies can reveal decision-making patterns without proving what happens when the same doctors face live patients, competing demands, and incomplete histories.
Even so, the design remains useful. It isolates the consultation logic around a biomarker that is already widely discussed in primary care. In doing so, it frames a question that stewardship programs cannot ignore: whether CRP serves mainly as a helpful guardrail against unnecessary antibiotics, or whether it becomes another routine test that clinicians reach for before deciding that a prescription is needed.
The abstract leaves that downstream clinical effect unresolved. What it does establish is more basic and more revealing: in routine decision-making, Norwegian GPs had far greater readiness to request CRP than UK GPs, and that single difference may shape how antibiotic decisions are made long before the prescription pad comes out.