DR. PETER CHOW: ‘Over-reliance’ on technology a threat to the ‘humanistic art and diagnostic science’ of medicine

Article content

The Groucho Marx Law of Medicine – Interpreting Medical Tests – A Test Can Only Be Interpreted in the Context of Prior Knowledge.

Article content

This seems, on the surface, to be a rule taken from the Groucho Marx handbook. (“I refuse to join any club that would have me for a member.”)  You need to have a glimpse of an answer before you have a glimpse of the answer.

To understand the logic behind this paradox, we need to understand that every test in medicine – any test in any field, for that matter, has a sensitivity, a false-negative rate and a specificity, a false-positive rate.

Sensitivity refers to a test’s ability to designate an individual with disease as positive.

A highly sensitive test means that there are few false-negative results, and thus fewer cases of disease that are missed.

The specificity of a test is its ability to recognize an individual who does not have a disease as negative.

Article content

A highly specific test means that there are few false-positive results, fewer cases of people wrongly identified as positive.

The point is that if patients are screened without any prior knowledge about their risks, then the false-positive and false-negative rates can frustrate any attempt at diagnosis.

Consider the following scenario. Suppose that the test for HIV has a false-positive rate of one in 1,000 –  i.e. one out of every 1,000 patients tests positive, even though the patient has no HIV infection. And, suppose, further, that we use this test in a population of patients where the prevalence of HIV is also one in 1,000.

A false-positive rate of one in 1,000 – most people would assume that a positive test would have a 999/1,000 – a 99.9 per cent chance of being accurate. But, for every one patient in 1,000 identified by the test as being infected with HIV, there will be one uninfected patient who will also test positive. In short, for every test that comes back positive, there is only a 50 per cent chance that the patient actually has HIV. We would all agree that such a test is worthless – it is right only half the time.

Article content

If the false-positive rate rises to one per cent, one in 100, 20 in 2,000 and the prevalence falls to 0.05, one in 2,000 – both realistic numbers with HIV – then the chance of a positive test being accurate falls to one in 21, an abysmal five per cent. The test is now wrong 95 per cent of the time.

In actual fact, the incidence rate of HIV for the overall Canadian population for 2024 is 4.7 per 100,000 population. The false-positive HIV ELISA test rate is 1.3 per cent – 1,300 in 100,000 will be false-positive.

So, a positive test has a 4.7/1,300 – a 0.36 per cent chance of being right – wrong 99.64 per cent of the time.

A totally useless test. In contrast, watch what happens if, on taking a detailed medical history and doing a thorough functional enquiry, we learn that the patient – sporting multiple tattoos and body piercings – is Indigenous, gay and an IV drug user.  We learn that the risk of HIV in gay IV drug users is 19 per cent.

Article content

The prevalence of HIV in Indigenous people is 3.8 times greater than in the general population, Gladue factors notwithstanding.

(In a study in the U.S., one in six gay and bisexual men will be diagnosed with HIV in their lifetime, including one in two Black gay and bisexual men, one in four Latino gay and bisexual men, and one in 11 white gay and bisexual men. Globally one in nine IV drug users will have HIV.)

Now, the up-front chance of HIV infection climbs to 24 in 100, while the false-positive rate is still one per cent, one in 100 tests.

Now, for every 25 positive tests, only one is a false positive and 24 are true positives – an accuracy rate of 96 per cent. The same test is transformed from being perfectly useless to being perfectly valuable.

Article content

All that is needed is a strong bit of “prior knowledge” applied in the preselection process to overcome the weakness of a test.

The “prior knowledge” is the kind of thing old-school doctors do very well, and that new technologies in medicine often neglect.

A thorough “Functional Enquiry” – the series of seemingly irrelevant systems-related questions used to elicit new information and further details about the presenting problem

“Prior knowledge” is at stake when your doctor, rather than just ordering another diagnostic lab or diagnostic imaging, asks you about your grandparents, whether you’ve had any exertional dyspnea, swelling of your ankles, PND (Paroxysmal Nocturnal Dyspnea), palpitations or checks your pulse and listens to your heart sounds for no apparent reason.

Article content

Your doctor is doing the thing that incisive doctors do: he is weighing evidence and making inferences. He is playing with probability, changing the odds. And yes, we profile patients in the ER and in medicine – the Tattoo Sign and the Body Piercing Sign in this patient signal probable membership in a specific badass subculture.

We make assumptions regarding our patients, their health status, and their risk factors based on their demographics or appearance – in other words, engaging in bias – not to denigrate them but to optimize their medical outcome.

(For those badasses in that particular subculture, denigrate means to “put down.”)

The medical profile of a patient is home to important patient details including their allergies, problems, medical history, family history, social history and medications.

Article content

Patient profiling, an essential practice in health care, aims to enhance the personalization of care and improve overall patient outcomes, helping health-care providers understand the diverse needs and experiences of their patients.

Applying “prior knowledge” in the preselection process to overcome the weakness of a test is not a feature unique only to medicine.

It applies to any other discipline predicated on predictions, economics or banking, gambling or astrology.  The core logic holds true whether you are trying to forecast tomorrow’s weather or seeking to predict rises and falls in the stock market. It is a universal feature of all tests.

Medical tests must be interpreted in the context of prior knowledge of risk and prevalence. A medical test is not a Delphic oracle, a predictor of perfect truths.

Article content

For that, consult California Psychics or Benny Hinn. Rather, a medical test is a device that modifies probabilities. It takes information in and puts information out.

We feed it “input probability” and it gives us “output probability.” Garbage in, garbage out.

Take PSA testing.  Prostate cancer is a strongly age-related cancer. If you test every man over 40 with a PSA test, it is certain that the number of false-positives will overwhelm the number of true-positives. Only one in 350 men under the age of 50 years are diagnosed with prostate cancer – just under three men in 1,000

PSA has a false positive rate of about 70 per cent – 700 in 1,000. The PSA test is now wrong 700/703 – 99.6 per cent of the time.

False-positive test results are common with PSA screening; only about 25 per cent of people who have a prostate biopsy due to an elevated PSA level of between four and 10 ng/mL are found to have prostate cancer when a biopsy is done.

Article content

Thousands of needless needle biopsies and confirmatory tests will be performed, each consuming medical resources, adding pain, complications, anxiety, frustration and cost. Transrectal ultrasound-guided needle biopsy of the prostate is as exquisitely painful as it sounds. If you do the same test on men at higher risk, the results become somewhat more accurate, but the false-positive and false-negative rates are still forbidding.

Men who are at higher risk of prostate cancer include men over 60, Black men, men with genetic variants in BRCA2 (and to a lesser extent, BRCA1), and men whose father or brother had prostate cancer.

Add more data – family history, obesity, sedentary lifestyle, urinary symptoms, digital rectal examination – and the probability of a truly useful test keeps getting refined. Yet demands for indiscriminate PSA testing to “screen” for prostate cancer keep erupting from patients and advocacy groups.

Article content

Ebola is flaring up again in the Congo and west Africa. Assume that the test for Ebola picks up 98 per cent of patients infected with Ebola but has a four per cent false-positive rate – 400/10,000. Also, assume that 0.01 per cent of travellers from Africa actually carry the virus – 1/10,000. Should we screen all passengers flying in from Africa?

The chance of a positive Ebola test being real is 1/401 – 0.25 per cent; 99.75 per cent of positive tests will be false and health care will be overwhelmed hunting for the one case out of 401 that is real.

The “screen everyone for everything” approach, the hand-held all-body scanner in Star Trek, works if you have infinite resources and time, but begins to fail when resources and time are finite. The TV show Star Trek had a fictional Dr. Leonard McCoy, who used a device called a tricorder to examine patients in an instant. The tricorder has a detachable, high-resolution, hand-held scanner that can check all vital organ functions, detect the presence of dangerous organisms and diagnose medical conditions within seconds.

Article content

Sort of like that snazzy little electronic miracle device at your friendly neighbourhood naturopath. The fictional device has spawned a search for its real-life equivalent. Whole-body MRI scans have been around for more than a decade, but recently they have taken over social media, promoted by the airhead likes of Kim Kardashian and Paris Hilton.
On Instagram in 2023, Kim Kardashian, she of labia-shaming fame – “mine are small and cute” (TMI) – wrote about being scanned by what she called a “lifesaving machine.”

“It has really saved some of my friends’ lives and I just wanted to share,” she posted.

Paris Hilton shared a similar experience of having a full-body scan to “proactively absolutely” rule out breast cancer.

Article content

“I encourage every single one of you to go get a scan,” she said.

Social media influencers, celebrities and entrepreneurs, promote full-body MRIs as “super scans” that can detect early cancers and aneurysms in their earliest stages, before symptoms arise, alerting people to “silent killers.”

They’re promoted by radiology companies and private MRI clinics as a risk-free way of getting a “complete health check” to prolong your life. There are now Ontario private MRI clinics in Mississauga and Ajax-Pickering offering whole-body MRI scans to anyone who wants one – even if they have no symptoms whatsoever nor a referral from a doctor.

The implication is that almost everyone is a good candidate. The male package includes a Whole Body MRI, Prostate MRI Screening and Coronary Artery CT Calcium Scan (to assess likelihood for developing a heart attack) while the female package includes Whole Body MRI, Breast MRI Screening and Coronary Artery CT Calcium Scan … each for only $6,250 … an absolute bargain for a little peace of mind. OHIP or private insurance companies do not cover the cost of these tests nor are there any plans in the foreseeable future.

Article content

Full-body MRI scans do have unintended negative consequences.

The main concern with full-body MRI scans is that they can’t detect the earliest signs of disease, including cancers and dementia, but can pick up inconsequential findings leading to unnecessary diagnoses, mega-anxiety and treatments leading to individual harm and overloading the health system. Normal findings could leave you with a false sense of reassurance, as many diseases cannot be diagnosed in the early stages using these technologies.

This may cause you to ignore symptoms of illness in the future, which in turn might delay treatment.
Suspicious findings can be abnormalities that are not harmful to your health that lead you to more tests, possibly including the use of surgery, drugs, anesthesia and radiation. These tests, treatments and procedures, which are likely unnecessary, could put you at further risk.

Article content

There is no evidence that these highly sensitive scans provide any overall benefit for people at average risk of disease. They do offer a good risk of overdiagnosis, unnecessary anxiety, and a depleted bank account. Unfortunately, since graduation from medical school 50 years ago, I have witnessed the continual devaluation of the cognitive skills of medicine.

The thorough history, physical examination, and thoughtfully narrowed differential diagnosis have been neglected in favour of more diagnostic tests, more imaging and more specialist referrals. Over-reliance on algorithms, the smart phone with Google, the advent of artificial intelligence, ChatGPT, and the utilization of non-physician practitioners as primary diagnosticians beyond the scope of their training (or lack thereof), are threats to the humanistic art and diagnostic science of medicine.

Article content

Perhaps, in the future, we can imagine a doctor who doesn’t have to take a careful history, ask seemingly irrelevant questions about your ancestors, listen to your heart sounds, do a rectal examination or watch the rhythm of your gait as you walk into the office. Perhaps all the uncertain, unquantifiable priors and inferences will become obsolete. By then, medicine will have changed.

We will have to learn new Laws for Medicine.

“These are my principles, and if you don’t like them … well, I have others” – Groucho Marx

Groucho’s last words: ” Die? … that’s the last thing I’ll do. This is no way to live.”

Dr. Peter Chow is a retired Sault Ste. Marie physician 

Article content