The risks of medical investigation are often overlooked

Jack Roberts is a salaried GP at Hammersmith and Fulham Partnership with interests in clinical research and medical education.


All medical interventions come with a degree of risk: medications have side effects, procedures have complications. As a result, it is standard practice for doctors to raise awareness of the potential pitfalls before patients choose whether, for example, to start a new drug or undergo surgery. This principle, of facilitating and observing informed choice, constitutes one of the pillars of ethical medicine.1

But what about investigations: blood tests, samples and scans? Are they any different to other medical interventions? On the surface, investigations can seem like pure, benevolent aides to both doctor and patient: they help us diagnose, and they shine light on a way forwards. However, dig a little deeper and it is clear that investigations too are fraught with risk. The technical risks of investigations are well-documented, and these are often flagged in doctor-patient interactions: bruising following a blood test, exposure to radiation during a CT scan. However, equally important, and regularly overlooked, are the risks associated with the essence of investigation itself. “Seek, and ye shall find”: but in many cases what we find generates far more questions than answers.

This principle, of facilitating and observing informed choice, constitutes one of the pillars of ethical medicine.

Consider a young man with weeks of generalised headaches. He works in a restaurant, a stressful job with unsocial hours. He is regularly dehydrated. He does not wear the glasses his optician has suggested when at work. Nevertheless, he is worried about a brain tumour, and would like a head scan. In an attempt to reassure him, his GP arranges an MRI. The scan shows a normal brain, but there is a small indeterminate nodule in the frontal sinus. The patient is understandably worried about this, and agrees to a referral to an ENT specialist for further assessment. The specialist arranges a biopsy, and the histology confirms a benign polyp. Two weeks down the line, the patient develops sinusitis, a complication of his biopsy. Several courses of antibiotics later, the sinusitis persists. The tension rises and his headaches worsen. Meanwhile, he continues to drink poorly, and still does not wear his glasses.

Now, consider a woman with tiredness. She works a 40-hour week and is a single mother to three children. She drinks half a bottle of wine every night. The GP thinks her tiredness is related to her life circumstances, but she requests blood tests. The GP reluctantly agrees to arrange them. The test results return entirely normal, aside from a slightly raised MCV; this is seen again in a repeat test 3 months later. Although her GP attributes this finding to alcohol, the patient is not convinced: she reads online that this can be an indicator of some form of blood cancer, and becomes quite fixated on a serious diagnosis. The GP struggles to reassure her despite numerous further investigations and specialist opinions. The patient develops anxiety, struggles to sleep and starts missing work.

These examples might be fictional, but they reflect cases we encounter daily in the world of general practice. There are lessons to be learned from them. First of all, they demonstrate that the decision to investigate a particular problem should not be taken lightly. Just because an investigation exists, it does not necessarily need to be used: indeed, if the degree of clinical concern is low, the risks of medical investigation may significantly outweigh the benefits.  Furthermore, such cases also highlight that patients ought to be made aware of the risks of investigation itself in the process of gaining informed consent: the dilemmas that “abnormal” results can present when there is a low index of suspicion, and the anxieties caused by unveiling incidental abnormalities of uncertain significance.

Curiously, there is a small, particular collection of investigations for whom we do emphasize such risks; investigations for which we laboriously highlight the “cons” with patients, alongside the “pros”. Take PSA, for example. As a GP, I can expect to discuss PSA testing on a near-daily basis, flagging the numerous associated practical, technical and ethical challenges that it presents2. Similarly, at medical school, I remember numerous communications workshops focused on counselling for HIV testing: the myriad implications of a positive test, the pseudo-assurances of a negative test. In antenatal settings, mothers-to-be will be advised on the pearls and pitfalls of screening for possible chromosomal disorders in the developing fetus, and the dilemmas that may be presented by a “positive” result. On the other hand, we doctors rarely mention the potential fallibilities of investigation when we propose a faecal calprotectin, an ultrasound of the renal tract, or an MRI brain. Instead, we tend to arrange them, and when the results prompt further questions, we counsel in retrospect. There is clearly a lack of consistency in our approach.

If we do not highlight that investigations may present results that are almost impossible to interpret without further tests, or that they risk revealing incidental changes of uncertain significance, can we actually say that we have gained true, valid, and informed consent to arrange them? Further still, is it even morally right that we can arrange tests that reveal deeply personal information about patients’ bodies, beyond that which is relevant to their presenting problem, again without their explicit permission? Imagine that a female patient has a CT thorax, abdomen and pelvis for unexplained weight loss, and the scan shows that she has bilateral breast implants. She has never disclosed this to anybody and was not warned ahead of providing consent that the scan might also unveil such incidental findings. When she reads the radiologist’s report, she is horrified. Does she have a right to be? This needs a professional consensus: in a world increasingly concerned by unsolicited sharing of personal data, the implications are significant: as discussed in the introduction, facilitating informed consent is a cornerstone of medical practice, and we cannot be doing a half-job.

…the investigation process should not be initiated without careful thought: without it, investigations themselves can morph from friends into foes.

I am aware that, so far at least, I have presented the process of “medical investigation” in a negative light. I feel it important at this stage to highlight that I am by no means “anti-investigation”: despite what traditionalists might say about history and examination, it is clear that well-utilised, targeted investigations are our most potent diagnostic tools, and to use them gives us our greatest chance of getting an accurate medical diagnosis. Further still, incidental findings can sometimes be hugely beneficial, and such discoveries can even save lives: there are countless examples in which asymptomatic haematological derangements, early-stage cancers, and advanced aneurysms have been found purely “by the grace of God” during the work-up for completely unrelated medical problems. Had those patients refused their respective investigations on the basis that might identify abnormalities that could create additional questions or dilemmas, many might have suffered tragically premature deaths.

However, this does not detract from three key points. First, the investigation process should not be initiated without careful thought: without it, investigations themselves can morph from friends into foes. Secondly, patients deserve to make fully-informed choices. Currently, when it comes to investigation, many are not given a platform to do so: they undergo it blind to the intrinsic risks, seeing investigations themselves as infallible no-brainers: “An MRI scan? Sure, why not?”. And thirdly: although the unexpected finding of something serious but curable might make for an uplifting story, we should not let it blind us to countless other examples of uninterpretable findings which result in totally unnecessary angst, worry and anxiety. To put it another way: the chance discovery of an actual early-stage, asymptomatic cancer outside a dedicated screening programme is rare; on the other hand, adrenal incidentalomas, lung nodules and renal cysts are common, and once found it is hard to ignore them, however likely they are to be completely benign.

This leads to one final argument. When we choose to investigate, we often start a process by which investigations beget investigations: blood tests lead to scans, scans lead to more detailed scans, and more detailed scans lead to endoscopies and surgical procedures. This has huge implications, not just for our patients, but for our health system as a whole. If we investigate without genuine clinical concern, be it to curb our own assurances or simply to placate our patients, we risk triggering a sequence of events that ultimately detracts from NHS resources for no justifiable reason. This live transition towards defensive and perhaps even consumeristic medicine undoubtedly contributes to our service’s gargantuan waiting lists and waiting times. This in turn leads to a disservice for those who are truly suffering, who desperately need investigations and specialist opinions so they can take the next step towards potential recovery.

In my opinion, we have a duty to raise greater awareness of the risks of “investigation” as a general concept. Investigations themselves have innate flaws: they can throw up results that are difficult to interpret, and they may reveal incidental, unrelated abnormalities of uncertain significance. To highlight these issues to our patients would constitute more ethical practice, would foster greater patient empowerment, and could result in a frame shift in the way our population approaches healthcare. Furthermore, and perhaps most importantly, doing so has the potential to significantly reduce inadvertent iatrogenic harm, whilst also alleviating the incredible burden on our health system.


  1. [accessed 19/3/24]
  2. [accessed 19/3/24]

Featured photo by Markus Spiske on Unsplash

Notify of

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

Inline Feedbacks
View all comments
Previous Story

At the National Theatre: Nye

Next Story

The Diary of a Somebody: on the banality of heroism

Latest from Clinical

Would love your thoughts, please comment.x
Skip to toolbar