Accepting and Managing Diagnostic Uncertainty

Medicine has evolved significantly over the past few decades. Technological progress has greatly improved physicians’ ability to establish diagnoses and provide increasingly less invasive treatments. Consider, for example, the advances made in cardiology and cardiac surgery over the last 20 years—and how many patients today benefit from procedures such as transcatheter aortic valve implantation (TAVI). Only a few years ago, some of these patients would have been considered inoperable for one reason or another.

However, this increase in medical certainty has made the uncertainty that exists in the absence of these technological wonders all the more visible. We can all collect a blood sample and measure troponin levels, giving us a relatively high degree of confidence when assessing a possible myocardial infarction. But how many of us could recognise a posterior myocardial infarction in its early stages based solely on an ECG and a clinical presentation dominated by dyspnoea? If the patient is diabetic, many general practitioners may find themselves confronted with precisely this problem.

In general practice, diagnostic uncertainty arises partly from the limited availability of complex investigations, most of which are predominantly accessible in hospitals. Another challenge is the way patients’ symptoms present. Every cardiology textbook mentions that an acute myocardial infarction may present solely as epigastric pain. But how realistic is it for general practitioners to perform an ECG on every patient presenting with symptoms suggestive of gastritis?

Even if we did, a new problem would emerge, further feeding the chain of uncertainty: an ECG alone cannot reliably exclude myocardial infarction. What should we do, then? Measure troponin levels in every patient? If so, how many times—once at presentation and again six hours later? If we attempted to solve diagnostic uncertainty simply by expanding the range of investigations, the healthcare system would eventually collapse under the resulting financial and logistical burden.

Another relatively modern factor contributing to the problem is the worldwide shortage of physicians. I cannot speak with confidence about Asia or the Americas, but most European countries are experiencing a clear shortage of healthcare professionals. Fewer doctors are being trained than in previous decades, many of those currently practising are approaching retirement, and Europe’s ageing population requires an increasing amount of medical care. It is the same imbalance between supply and demand that I discussed in another post about inflation.

Can we simply withdraw healthcare from part of the population? No—certainly not. Patients, like all of us, have a constitutional right to healthcare, just as we, as physicians, have a constitutional right to be protected by the police and other public institutions. The fact that the current situation is critical is not an excuse to abandon our ethical principles.

The problem arises when we try to care for patients while being undermined by a healthcare system in free fall. There simply is not enough time for everyone. Insufficient time is yet another source of uncertainty in clinical practice.

Diagnostic confidence can be improved when we are able to recognise cases that may deteriorate and devote sufficient time to them. Similarly, combining several tests may improve diagnostic accuracy, although no test is perfectly accurate and the limitations of each must be thoroughly understood.

When we cannot further improve patient safety or diagnostic accuracy, it is essential to establish a clear safety-netting plan with the patient. This plan should include precise advice and clearly defined conditions for seeking further medical attention.

For example, parents of a two-year-old child might be advised to attend hospital if they are unable to keep the child adequately hydrated or if a temperature of 39°C does not improve within the next six hours. In this context, however, “adequately hydrated” is far too vague for someone without medical training.

A clearer instruction would be to seek medical attention if the child has not had a wet nappy during the following six hours. This gives the parents a concrete warning sign that connects reduced fluid intake with decreased urine output and possible dehydration.

Most patients respond positively when diagnostic uncertainty is communicated openly, when the reasons behind it are explained, and when they are given a clear safety-netting plan.