Uncertainty in Medicine
One of the reasons I went into medicine was the naïve idea that doctors always know what to do. I was wrong. Marya Zilberberg got it right when she said, “The only certain thing about medicine is uncertainty.” Patient histories are uncertain, physical exams are uncertain, tests are uncertain, diagnoses are uncertain, treatments are uncertain, even human anatomy is uncertain. Doctors are not scientists; they are practical users of science who apply scientific evidence to patient care. Medicine deals in probabilities and informed guesses, not certainties. Symptoms can mean many things. They can be a sign of serious illness requiring treatment, an illness we don’t yet know how to identify and treat, a benign condition that will resolve without treatment, a hyper-awareness of normal bodily functions, depression, somatization disorder, malingering, or a cry for help.
The patient’s account of his symptoms and medical history is uncertain. Memory is unreliable. We forget, we distort, and we remember things that never happened. Patients want to present themselves in a good light; they tend to underestimate their alcohol and tobacco use and hesitate to mention sexual indiscretions and foolish use of nonsensical remedies. Their answers to “how much pain on a scale of 1 to 10” depend on variables such as their current mood and their degree of stoicism.
Doctors do a “review of systems” (often by a printed questionnaire) asking about symptoms in different body systems (respiratory, gastrointestinal, etc.). Answers will depend on how significant the patient thinks the symptoms are. One patient may answer “no” to heartburn because he thinks his heartburn is not worth mentioning; another may say “yes” because he remembers he had a mild burning sensation for about 2 minutes once last month.
Physical exams are uncertain. It’s easy to miss an abnormality like an enlarged spleen unless you suspect it and try hard to find it. Even anatomy is uncertain. Some people have reversed organs (situs inversus), where the heart is on the right and the appendix on the left. Some are missing a kidney. The location of blood vessels and nerves is variable. Look at the veins on the back of your hands; they’re not in exactly the same place on each hand.
Even something as apparently straightforward as the obstetrician’s announcement “It’s a boy!” can be wrong. In my recent article on gender differences (“Gender Differences: What Science Says and Why It’s Mostly Wrong” in Skeptic 18.2) I listed five determinants of biological sex in a newborn (sex chromosomes, gonads, hormones, internal genitalia, and external genitalia) and the many ways in which each of these can go wrong. Other factors later in life contribute to classification of sex and gender: development of secondary sexual characteristics at puberty, the sex the person was reared as, gender self-identification, object of desire, behavior, dress, role in society, and legal gender.
Laboratory tests are uncertain. Normal lab values are determined by testing lots of normal people, creating a bell curve, lopping off both extremes, and arbitrarily designating the central 95% as normal. Results outside the normal laboratory range aren’t necessarily abnormal, and results within the range aren’t necessarily normal. If a healthy person gets 20 tests, one is likely to be outside the “normal” range just by chance. Lab errors can occur from mislabeling, failure to calibrate instruments, procedural errors, and clerical errors. Diurnal variation, pregnancy, drugs, diet, exercise, and supplements may affect test results.
Imaging studies can be misleading or misinterpreted. A famous fMRI of a salmon appeared to show that it was thinking, but the salmon was dead. When the figure of a tiny gorilla was added to a chest CT scan, 83% of radiologists who read the x-rays looking for signs of lung cancer missed the gorilla: an example of inattentional blindness.
Screening tests involve a lot of uncertainty. False positive results can lead to unnecessary worry and further investigation including invasive tests that may result in harm or death. The United States Preventive Services Task Force (USPSTF) recommends the following tests not be used for routine screening because they are likely to do more harm than good: PSA tests, EKGs, TB tine tests, scoliosis checks, and chest x-rays.
Doctors are not scientists; they are practical users of science who apply scientific evidence to patient care. Medicine deals in probabilities and informed guesses, not certainties.
Direct-to-consumer genetic testing can be misleading. A blue-eyed man was told he had the genes for brown eyes. Testers only look for specific SNPs (single nucleotide polymorphisms) and report probabilities based on imperfect information. They may report that people with your SNP are 30% more likely to develop Parkinson’s disease than people with other SNPs. But disease is not destiny. Even if you have the gene for a disease, that gene may or may not be expressed. Gene expression depends on environmental and epigenetic factors and on interactions with other genes. Our access to genetic information currently exceeds our understanding of what that information actually means.
Diagnosis is uncertain. Some conditions, like Alzheimer’s disease, can’t be definitively diagnosed until autopsy. Tests don’t make a diagnosis; they only raise or lower the likelihood of the diagnosis compared to the pre-test likelihood. The Diagnostic and Statistical Manual of Mental Diseases (DSM) has been widely criticized: its reliability and validity have not been established, it categorizes by symptoms rather than causes, it is culturally biased, it tends to medicalize common life problems, and it labels and stigmatizes patients. Earlier editions of the DSM labeled homosexuality as a psychiatric disorder.
Doctors must constantly make decisions based on inadequate information. They can’t afford “analysis paralysis.” If they hesitate like Hamlet, their patients may die. They walk a fine line between indecisive humility and the reckless arrogance of overconfidence. Surgeons in particular must decide and act swiftly before the patient bleeds out.
Will Rogers said, “It isn’t what we don’t know that gives us trouble, it’s what we know that ain’t so.” For years we “knew” that routine episiotomies, radical mastectomies, and internal mammary artery ligations were benefitting our patients; but controlled studies showed they weren’t, so we stopped doing them. For centuries doctors were sure bloodletting worked, but it was actually killing patients.
Treatment involves a lot of uncertainty. Choosing the best drug for each patient is mostly a crapshoot. Genetic testing promises to eventually guide us in individualizing drug choices, but it isn’t of much practical use yet. Choosing an antibiotic for pneumonia or meningitis is a guessing game based on exposure, travel, immune status, allergies, the most common causes of those infections, and the prevalence of bacterial strains in the community. Waiting for culture and sensitivity testing is not an option, since the patient may die before the results are back.
There is uncertainty in communication. Patients don’t remember most of what a doctor tells them. In a recent study of patients on palliative chemotherapy for terminal cancer, many patients were under the false impression that it offered them a chance of cure or longer survival when its only purpose was to ease their dying.
When no diagnosis is found, doctors and patients must decide when to stop testing and accept uncertainty. Once serious diseases have been ruled out, the likelihood that further tests will identify something significant and treatable is diminishingly small. Too much testing can hurt the patient through false positives, unnecessary anxiety, false hope, wild goose chases, and invasive, dangerous, expensive diagnostic procedures. At some point we should stop asking “why” and focus on “how” to cope with symptoms and improve quality of life. Three things can happen. The symptoms may go away (in which case, who cares what caused them?). They may stay the same, in which case we can keep trying to find better ways of coping. Or they may get worse, in which case we can always reevaluate and reconsider the need for further tests.
We can learn to live with uncertainty, and surely it’s better to be uncertain than to be certain and wrong. Voltaire said, “Uncertainty is an uncomfortable position. But certainty is an absurd one.”
CAM (complementary and alternative medicine) deals in certainties. The chiropractor is certain your symptoms are caused by subluxations and treatable with manipulations. Naturopath Hulda Clark was certain all disease was due to parasites, and was certain she could cure everything with her zapper; but she failed to cure herself. Sherry Rogers, author of Detoxify or Die, is sure diseases are due to toxins. A business man named Robert Young is sure they are due to acidosis. Acupuncturists are sure disease is due to obstructions or imbalances in the flow of qi.
I once searched the web for “the one true cause of all disease” and found 67 of them! They included refined sugar, grains in the diet, stress, ama due to aggravated doshas, free radicals, sin, allergies, poisonous chemicals, a congested colon, witchcraft, overeating, food acidity, toxic metals, arrogance, impairment of movement of the bones of the skull, cold, and poverty. On one website I learned that “The United KKK States of America is the root cause of all disease.” I have no idea what that means, but I guess it’s convenient to have someone to blame.
Cranks and quacks lack humility in the face of disease; they demonstrate the arrogance of ignorance. They lure their victims with false promises of miracle cures for incurable diseases and less scary “natural” alternatives to surgery, chemotherapy, radiation, and drugs. They don’t know what they don’t know, and that makes them very dangerous.
Modern medicine is riddled with uncertainty, but it’s still far better than any other option. Doctors understand basic science, realistically judge probabilities, wrestle with the realities of uncertainty, and make informed guesses based on the best currently available evidence. CAM deals in certainties based on fantasy and intuition. Scientific medicine progresses over time and discards treatments that don’t work; CAM never admits errors and makes no progress.
Uncertainty can be a good thing.