The first population study in history was born out of a dramatic debate involving leeches, “medical vampires,” professional rivalries, murder accusations, and, of course, bloodletting, all in the backdrop of the French Revolution. The second of a multipart series on the development of population medicine, this episode contextualizes Pierre Louis’ “numerical method,” his famous trial on bloodletting, and the birth of a new way for doctors to “know”. Plus a brand new #AdamAnswers about Occam’s razor and Hickam’s Dictum. All this and more on Episode 38 of Bedside Rounds, a tiny podcast about fascinating stories in clinical medicine! To claim CME and MOC credit, please go to www.acponline.org/BedsideRounds.
- Best M and Neuhauser D, “Pierre Charles Alexandre Louis: Master of the spirit of mathematical clinical science,” Qual Saf Health Care 2005;14:462–464.
- Duffin J, “Laennec and Broussias: The ‘Sympathetic’ Duel,” from La Berge A and Hannaway C, Paris Medicine: Perspective Past and Present.
- (1977) The French Revolution: A Revolution in Medicine, Too, Hospital Practice, 12:11, 127-138
- Hillard A, et al. “Occam’s Razor versus Saint’s Triad, N Engl J Med 2004;350:599-603.
- Lo Re V 3rd, Bellini LM, William of Occam and Occam’s razor. Ann Intern Med. 2002 Apr 16;136(8):634-5.
- Kirk GW and Pemberton N. Leech, 2013
- Kirk GW and Pemberton N, Re-imagining Bleeders: The Medical Leech in the Nineteenth Century Bloodletting Encounter. Med Hist. 2011 Jul; 55(3): 355–360.
- La Berge A and Hannaway C, Paris Medicine: Perspective Past and Present.
- Louis PCA. Researches On The Effects Of Bloodletting In Some Inflammatory Diseases. Boston: Hilliard, Gray, 1836.
- Morabia A. PCA Louis and the birth of clinical epidemiology. J Clin Epidemiol 1996;49: 1327-33
- Morabia A, Pierre-Charles-Alexandre Louis and the evaluation of bloodletting. J R Soc Med. 2006 Mar; 99(3): 158–160.
- Niehyl PH. The English bloodletting revolution, or modem medicine before 1950. Bull Hist Med 1977; 51, pp. 464-483.
- Papavramidou N and Christopolou-Aletra H, Medicinal use of leeches in the texts of ancient Greek, Roman and early Byzantine writers. Intern Med J. 2009 Sep;39(9):624-7.
- “Suckers for Success,” Nature volume 484, page 416 (26 April 2012).
- “Walter Chatton,” Stanford Encyclopedia of Philosophy, retrieved from: https://plato.stanford.edu/entries/walter-chatton/
- Wardrop D, “Ockham’s Razor: sharpen or re-sheathe?” J R Soc Med. 2008 Feb; 101(2): 50–51.
This is Adam Rodman, and you’re listening to Bedside Rounds, a tiny podcast about fascinating stories in clinical medicine, produced in partnership with the American College of Physicians. Yes, you heard that right! I’ve now partnered with the ACP; I have more details after the episode, but if you’re a member of the ACP you can now get CME and MOC points for listening to Bedside Rounds! Just go to www.acponline.org/bedsiderounds and take a brief quiz. This is part 2 of my series on the birth of population medicine, called Blood on the Tracks, and it’s about how growing controversies about bloodletting led a French physician named Pierre Louis to develop his “numerical method” to run the first population study in history. Along the way we’re going to talk about leeching, murder accusations, nosology, medical vampires, the French Revolution, and fundamental questions about how doctors can every really know how a therapy works. And at the end, a #AdamAnswers on Occam’s Razor and Hickam’s dictum!
That’s a lot for one podcast host, so to tell this story, I’m joined by my friend, colleague, and one of the smartest people I know, Dr. Shani Herzig. So first, a brief introduction:
20 secs: I am the director of hospital medicine research at Beth Israel Deaconess medical center and a hospitalist.
And what type of research do you do?
I do mostly something called pharmacoepidemiology, I focus on hospital-based outcomes and how we can improve care processes, a lot of medication safety work.
How big are the group of patients that you’re looking at?
I’ve done studies that have had over 3 million then some that have a few hundred. On average talking about a few hundred thousand
3:05 I think this is a nice tie in to what we are going to talk about, because in many ways you are a 21st century Pierre Louis.
I’m flattered. That might be the nicest thing anyone has ever said about me or my work. 7
But first I’ll do a TL;DR recap from episode 37. Bloodletting is an ancient therapy, but in the late 18th century, doctors started to remove more and more blood from sicker and sicker patients. Controversial cases in the new United States — Benjamin Rush’s plea to “venerate the scalpel” during the Philadelphia yellow fever outbreak of 1793 and the ensuing libel case, as well as the death of the first president George Washington led to increasing criticism of massive bloodletting. These medical debates around two key figures from the American Revolution were nasty, but nothing compared to the vitriol — and blood — that would be spilled in Revolutionary and post-Revolutionary France. America was very much a medical backwater at the turn on the 19th century.
Not so in France. After the French Revolution, starting in 1793, intellectual trends from the Enlightenment coalesced to violentally form brand new societal and governmental structures. Scientia was revered as a Goddess in Robespierre’s Cult of the Supreme Being, and church properties were confiscated and used to fund a new medical system. The revolution sought to, as philosopher and historian of medicine Michael Foucalt critically put it, organize “ a nationalized medical profession, organized like the clergy, and invested at the level of man’s bodily health, with powers similar to those exercised by the clergy over men’s souls, and the myth of a total disappearance of disease in an untroubled dispassionate society restored to its original state of health.”
And these new policymakers, many of them physicians themselves, set about to fundamentally change the practice of medicine. Their ultimate goal was the end of all disease and a society where hospitals would no longer be necessary — but they realized this couldn’t happen simply overnight. In the meantime, the country’s disparate medical schools were shuttered and replaced by three new schools in Paris, Montpellier, and Strasbourg. The old faculty was, for the most part, expelled, replaced by younger liberal physicians. Medical education was now free and open to all citizens. Education was reorganized from theory taught in the classroom to the bedside of crowded hospitals. To talk about all of the reforms of the French Revolution would take an episode in itself, so there’s a great article in the show notes outlining some of the more radical proposed and partially realized ones.
The Revolution is usually counted as the birth of the Paris School of medicine. I’ve talked about the Paris School quite a bit on Bedside Rounds — most recently in episode 34, about Corvisart, Laennec, Piorry, and the birth of the physical exam. Discussing their influence is difficult because there’s been so much hagiography; in many ways the Paris School is the foundational myth of modern medicine. The historian Edwin Ackerknecht, who has written probably the definitive account on the Paris school, identifies five intellectual trends that coalesced in the School: the rise of pathological anatomy; that is, performing autopsies on patients after they died; the incredible concentrations of sick patients in one place; the physical exam, often at the expense of listening to patients; the birth of “hospital medicine” — that is, the hospital as the fundamental place to practice and learn; and finally, as I’m going to talk about in this episode, the use of medical statistics.
Modern historiography of the Paris School, as you can imagine, is more complex than the traditional story. Briefly, many of these trends predated Paris, and modern analyses have suggested that rather than a clean break with the past, the Paris school should be seen as more of a continuation of 18th century medicine. And while modern accounts have tended to regard the School monolithically, there was in fact a huge variety of opinion and heated disagreement between its members. But the ideas from the Paris School made modern medicine, and they’re impossible to discount. Perhaps the influential idea emerging from this period is the change in the “clinical gaze”, or “le regard” as Foucault originally called it in French. Disease was now freed of metaphysical implications; sickness was not sin, and not part of inevitable human suffering. Disease could not only be treated — it could be eliminated.
The new revolutionary spirit needed a new revolutionary model of medicine — and it found it in Francois Joseph Victor Broussais. Like the Paris school in general, Broussais was a product of the French revolution. During the Terror, his parents had been massacred by Royalists, and forever after he was a committed liberal, atheist, and an iconoclast committed to tearing down the past, not only Hippocratic and Galenic medicine, but also the more recent pathologic anatomy practiced by other members of the French school. Like Benjamin Rush in Philadelphia, Broussais became frustrated with previous attempts to subdivide disease, including Cullen, and also his successor in France Pinel, who had attempted to merge many of the pathologic changes seen in autopsy with Cullens orders, genera, and species. And he was even more frustrated at the inability — and perceived hesitancy — of doctors to try and treat disease. He felt that other members of the Paris school were simply content to observe the sick and dying. Broussais even accused his fellow members of the French School, especially Laennec, of purposefully not treating their patients so that they could be autopsied after.
In 1808, he published his description of “physiological medicine,” which rapidly became the cause celebre of the Paris School. All disease has the same fundamental underlying cause, which is an inflammatory response to stress, originating as a gastroenteritis in the stomach. The specific symptoms of the disease depended on which organs the inflammation settled in. Continuing similarities with Benjamin Rush, Broussais did not believe in the healing power of nature — only by “heroic” means could the doctor save his patients. And its practice and rhetoric was heavily influenced by revolutionary politics; copious bloodletting relieved the patient from oppression, as opposed to the “medicine of symptoms” which treated nature as an autocrat which needed supportive therapy. Unlike Rush, however, Broussais believed that specific organs could be targeted with bloodletting — both systemic bloodletting by means of a lancet, but more identified with the use of leeches. Broussais advocated leeches being placed over the affected organ, which often meant the chest in this period of endemic tuberculosis. And not just one or two — a typical treatment might involve up to 90. Broussais has fallen into obscurity now, but at the time he was the most famous and respected member of the Paris school, and his influence was felt throughout Europe.. Widespread leeching of sick patients became widely practiced, including in England and the United States.
Leeching was not a new therapy and, like bloodletting itself , dates to ancient times. A mural that appears to show the application of leeches was found in any Egyptian tomb that dates to 1600 or so BCE in Egypt. By the time of Galen, their use had been well described, especially in diseases with an excess of blood, such as priapism or nasal congestion, where the leech would be applied to the offending organ. The English word “leech” maybe comes from an old-English word “leech” meaning physician — though the OED disagrees. But leeching was a small scale operation until Broussais came around. Take, for example, the poem Resolution and Independence by William Wordsworth in 1809. He describes a traveler walking among the moors who comes upon an old man. He asks what he’s up to:
He told, that to these waters he had come
To gather leeches, being old and poor:
Employment hazardous and wearisome!
And he had many hardships to endure:
From pond to pond he roamed, from moor to moor;
Housing, with God’s good help, by choice or chance;
And in this way he gained an honest maintenance.
The poem doesn’t go into the hardships he endured — but leech farmers would strip to their thighs and walk with their exposed legs through ponds, enticing the leeches to bite. Afterwards, these leeches would be sold to doctors. But with Broussais’ new ideas, the demand for leeches just couldn’t keep up with “artisanal” leech production. In 1833, for example, France imported 42 million leeches for medicinal purposes alone. At first, leech ponds were supplemented with old horses, whose legs would be cut, and then they would be run through the ponds. But there was still a leech shortage, which spurred the development of “leecheries”. Leecheries were artificial ponds, often located in cities, close to patients, where leeches could be farmed. But they quickly became associated with lower quality, and attracted birds, rats, and vermin. I want to quote an excellent piece by Kirk and Pemberton, of the damage these creatures could do to leecheries:
“In 1864, for example, a disaster was reported in a large French leechery caused by the arrival of a flock of wild geese. Presumably resting en route to warmer climates in what looked like an inviting new pool in the Parisian suburbs; these geese reportedly caused the loss of over 200,000 leeches in one incident (though it is not clear whether the leeches were consumed by the geese or absconded by affixing themselves to the bellies of the birds)”.
I think it’s also important to mention that in many ways leeching was a pleasant experience, at least compared with bleeding from the lancet. A “leech tube” would be used to place the leech directly on the affected area. You can see an image of these that I uploaded to Twitter. How much blood can leeches actually remove? The average amount of “bloodmeal” for hirudo medicinalis is just 2.5 mL over 2-4 hours, meaning 90 leeches remove about 225 mL of blood. Leeches became a fashion item; apparently at the height of leech mania, dresses and ribbons embroidered with leeches were in vogue; the style was called “les robe a la Broussais.”
Broussais had largely won in the court of public opinion, and patients flocked to man who was called the “Vampire of Medicine.” But in the medical community, his physiological medicine was incredibly controversial, especially with Rene Laennec. Laennec, of course, is probably the most famous member of the Paris School today — he invented the stethoscope and basically described the approach we take to diseases of the chest today. But in the early 1800s, Laennec and Broussais were locked in a decades-long duel. Some of it was personal; Laennec was a soft spoken Royalist and Catholic, which enraged the brash, liberal Broussais. But much of it came down to debates about nosology. Laennec and his supporters pointed out that such a simplistic system couldn’t explain the changes that they were seeing on autopsy. Broussais, in Laennec’s words, was “imprecise” and a poor pathological anatomist. “Inflammation” was more than a nebulous concept; there were real differences in gastric ulcers, in peritonitis, and inflammation of the lungs. But Broussais argued back that their view was abstract and fatalistic; even with these pathological findings, their treatments were essentially the same as two thousand years before — small amounts of bleeding, bland food, and purgatives, all designed to help the body naturally “balance” itself. Broussais’ theory, on the other hand, sought to actually treat the underlying cause.
So that’s our context; it’s the 1820s; physiologic medicine is in vogue. And a young doctor, very much on the periphery of the Parisian medical establishment, is pulled into the fray — and would forever change medicine in the process. Enter Pierre-Charles-Alexander Louis. Louis was as much a product of the Revolution as Broussais. He was the son of a wine merchant, and his father died when he was six. Under the Old Regime, his world would have been narrowed, but with the Revolution, a university education was opened to him. He settled on medicine and graduated in 1813. He then traveled to Russia to start a practice, eventually settling in Odessa. He had a thriving practice, and was given the honorary title of physician to the Tsar. But in 1820, a diptheria outbreak struck the port city.
Diptheria is one of those horrible diseases we’ve largely forgotten because of vaccination — the humble Tdap. It would start as a febrile, flu-like illness — but within a few days, a grey patch, called a pseudomembrane would start to form in the back of the throat, in the nose, and in the sinuses. The disease gets its name from this pseudomembrane; diphtheria comes from the Greek for leather. Patients with these dramatic pseudomembranes could be strangled to death from the inside — but most died from the toxin made from the bacteria, which can cause paralysis or abnormal heart rhythms.
One by one, Louis watched his patients die from the plague, his treatments powerless against the disease. I wish I could have known what was going on in his mind, but I suspect he was realizing some of the fundamental weaknesses of dealing with diseases that affected entire populations on a patient-by-patient basis. He closed his practice and moved back to Paris, where he took an unpaid job clerking la Piete Hospital, which soon turned into a job at Charity. But in Paris, he found that same fundamental weaknesses he had discovered in Russia, and he became convinced that a new method of approaching — and treating – – disease was necessary. In this way, he was not too different than Broussais. But Louis rejected such a simple nosology; he realized that a disease could not only be described by looking at individual patients. Louis would have to look at entire populations. He set out by making detailed notes on the patients that he cared for, like Laennec and the rest of the Paris school, including observations made during autopsy. But his innovation was to standardize and tabulate these observations as much as possi ble. He called this his “numerical method”. Ultimately, we don’t know where his influences came from. He was undoubtedly influenced by Laplace’s new ideas about probability, and Pinel, whose nosology had frustrated Broussais, had used similar methodology to argue for humane treatment of the mentally ill. Louis, like his Paris school colleagues, saw himself as an Empiric. His personal motto was “arts medica tota in observationibus” — the medical art consists entirely of observations.
But his numerical method allowed him to observe things that no one person could do on their own. He first made a name for himself by describing the effects of yellow fever on 123 of his patients, and he then repeated his methodology with typhoid fever. It’s unclear why he was sucked into the increasing debates about physiological medicine, and bloodletting as a whole, and for reasons unknown to us decided to use his numerological method to not just describe a population, but to test an intervention.
First caveat: it’s important to note that Louis was not looking to disprove bloodletting; he was specifically trying to challenge Broussais’ method. Like essentially every doctor in this period, he thought that bloodletting still worked, but in modest amounts, and in carefully selected patients. But how to do this? He found his answers in his copious records from Charity. He identified 77 patients who had been diagnosed with pneumonia and treated on his ward, but more importantly, had been completely healthy beforehand. I should note that his criteria were reviewed by a group of 46 chest physicians in the 21st century, who agreed with his classification; his meticulousness holds up to modern standards. Louis then calculated the onset of disease in each patient, and then analyzed the frequency of death grouped by first bloodletting. He decided to analyze them in two groups — an early group, which was bled on day 1-4 of their hospitalization, and a late blood letting group, bled on days 5-9. If Broussais were right, the first group should have a much higher survival. After all, they received the life-changing therapy earlier than the second, which should die at a higher rate.
But what Louis found showed him — he called it “startling and apparently absurd” — 44% of the patients who were bled early died, compared to only 25% of those in the later group.
Louis had made history, even if he didn’t realize it yet. What should we make of his findings? Again, Dr. Shani Herzig, modern day Pierre Louis:
3:42 I was actually just blown away by what he did even though it’s an incredibly primitive approach , you really start seeing the seeds of what modern day epidemiology has become. I think he would be pretty blown away by some of the techniques that we’re using nowadays to get at the things that he was alluding to, like confounding and bias. I was really, really impressed. Basically you can see in his writing that he is acknowledging confounding and making an attempt to actually negate or address it by virtue of assuring balance of different characteristics between his comparator groups which really is the essence of what confounding is. He even recognizes at the time that the systematic collection of data is very important and there are even hints in there that he understands that by ensuring that data is collected similarly for two groups that any misclassification will not be problematic because it will be nondifferential which is really an advanced concept that he’s already alluding to in his work.
He recognizes that selection bias and actually even immortal time bias, which is really a sophisticated concept, which could have played a role in his findings, which is to say that patients in the late bloodletting group needed to survive long enough to even make it into the late bloodletting group. …
Louis presented his conclusions in terms of simple arithmetic means — after all, like Shani pointed out, he didn’t have access to fancy statistical methods. But one hundred fifty years later, Dr. Morabia took Louis’ freely available data and converted them into something modern doctors understands — Kaplan-Meier curves, a type of survivorship curve, and he also posted the data freely available on the internet so you can play around with it. Both are naturally posted to my Twitter. But let’s do a little “graph on the radio.” Imagine two parallel lines. They each start at the top, where everyone is alive. But with each passing day, the each line drops a little more, showing that another patient has died. If you look at the survival curve between early and late bleeding, the early bleeding line plummets over the first two weeks, while the late bleeders drop far slower. For what it’s worth, Morabia calculated the p value between the two groups at 0.07.
Louis’s conclusions, first published in 1828, seem pretty conservative to us, but were provocative in their time. In fact, there’s evidence that he actually massaged his own data to minimize the effect, possibly worried his data would be otherwise rejected out of hand. The early group actually had a shorter duration of disease despite the higher mortality; with this fact he decided that bloodletting could be useful, though it was very dangerous, and should certainly not be given to everyone — a conservative strategy would likely be best for the majority of patients. Doctors should carefully decide which patients would most benefit — and once that has been decided, then massive amounts of bloodletting should be used.
This conclusions don’t quite hold up to modern scrutiny.
740: You know, it’s interesting. He only had access to people who had bloodletting and not people who didn’t have bloodletting so that’s kind of a fundamental difference between how I would have approached it and the way he approached it. But given what he had access to, he chose the right thing to do, which is to look at early versus late. I don’t think his work was rigorous enough to really draw firm conclusions. I think the thing that his work adds is really the methodologic approach as opposed to the findings. I think it would be impossible with the sample size he had to really meaningfully tease out confounding as an explanation for the findings. So I don’t take a whole lot from the study, like, I’m a physician, let’s not bloodlet. But I think more importantly it’s just the methodological advances that he initiated.
By this time, Louis was a respected member of the Paris school, and his paper did make a splash as part of the general backlash against Broussais in France. But Louis’ “numerical method” was not widely adopted. Doctors did not suddenly start analyzing patient data on large scales; they continued to publish individual case reports, or series of several patients. They continued to be empirics. But young people from around the world came to Paris to learn from the Paris School, including Louis. These students tak e back Louis’ fundamental message, that we can learn medical truths by studying populations and taking meticulous records, and would carry on the torch over their own careers — especially Oliver Wendell Holmes in the US, and William Farr in England. But that story — of Farr’s innovations in vital statistics, and the context for John Snow’s cholera studies — will have to wait for future episodes of this series, though I’ll warn you now, it might be a while.
So all this begs the question — does Louis’ stand up to modern scrutiny? Or has his biggest influence been the legacy of his students? The last hundred years have seen a lot of criticism of Louis and his famous paper, some of it fair, some of it not so much. He’s been criticized for his simple math, for a lack of originality in his approach, for lacking randomization. So I asked Shani what she would have done.
623: It’s hard to say what I would have done back then. Bak then I probably would have never done this to begin with because I wouldn’t have been smart enough to come up with it. But what I would do now to help determine the effects of bloodletting and whether or not it’s effective is obviously a randomized controlled trial. But I don’t do randomized controlled trials for the most part. You could design a really nice observational study around this using much more sophisticated methodology. I would do a large cohort study, use multivariable logistic regression, put a lot of different variables in my model, I might even get very fancy and do a propensity score approach, wherein I would take patients who had bloodletting done, and then match them with patients who have an equal propensity for bloodletting, but didn’t get it. You could washout any of the differences of those two groups. So that’s probably what I would have done.
Louis occupies an odd transition point in medical history. He’s usually recognized more as a “proto-epidemiologist” than for his work as a physician. But I don’t think that’s really fair — which, of course, would be my opinion since I’ve spent the last two episodes and about the last hour of your life building him up. Louis had a clinical question, and invented a potent new way to discover hidden truths.
1020: It’s moving from empiricism to, kind of, this goes beyond what an individual physician can see in their lifetime. I think that idea, there are still times where physicians are anchoring to their own personal experiences. We still talk all the time about different cognitive biases and their effect on clinical decision making. Physicians of older generations are still very reticent to accept evidence based practice.
If you pay attention in the hospital, you can hear this tension between our three different ways of knowing. Instead of “Empirics,” we talk about clinical judgement. Instead of experimental physiology, we have all sorts of medical sciences that our ancestors couldn’t have imagined — genetics, molecular biology, organic chemistry. And instead of a numerical method, we have randomized-controlled trials, retrospective cohort trials, and of course observational studies with propensity score matching. I think most doctors understand that there’s some balance between these three approaches to knowing. But the fundamental tension introduced by Louis– do we know by seeing our patients, by doing experiments, or by looking at populations — is if anything stronger today than it was in the 1830s.
2250 That was what was really resounding when reading all of this. It really is having humility in our field. And that it was I love about Pierre’s approach and what I do. Quote from Donald Rumself GOES here. Unknown unknowns — latter category proves problematic. That quote is actually so highly relevant to the work that I do. Really playd through my head in everything I was reading about Pierre Louls. Even w hen we think we know something, our entire foundational knowledge COULD be wrong. Could be something that shifts it on its head. Realizing that is one of the most important things we do as physicians. I think that population medicine really comes from a position that we don ’t know a lot of stuff, and when we do know it we need to test it.
With that in mind, what happened to bloodletting? It took a long time to finally die. Louis and his contemporaries broadsides worked, and over the remaining 19th century, its use slowly dwindled. In 1855, a doctor influenced by Louis, William Bennet, used mortality statistics to show that the “restorative principle” had a dramatically lower death rate in pneumonia than bloodletting — one third of the bled patients died, compared to about four percent of the non-bled. Then germ theory gave a new explanatory model for “inflammation” — infectious microorganisms — that destroyed the intellectual basis of bloodletting. But even with all this, bloodletting managed to linger. William Osler addressed it in his 1892 textbook, writing, “During the first 5 decades of this century the profession bled too much, but during the last 5 decades we have certainly bled too little.” And what did Osler think the chief indication was for bloodletting? Early pneumonia. The final 1942 edition — yes, 1942 — edited by McCrae, includes this line: “To bleed at the onset (of pneumonia) in robust healthy individuals in whom the disease sets in with great intensity and high fever is good practice.”
What was Osler’s reasoning? Likely his own observation and robust clinical experience. Which is just another way of saying that the more things change, the more they stay the same.
Well, that’s it for the show! But stay tuned after #AdamAnswers, because Shani and my conversation will continue for those interested — about how we, as 21st century physicians, use empiricism versus population medicine on a daily basis, whether there’s a paradigm shift afoot to give us doctors a new way of “knowing,” and finally about her favorite study.
And now it’s time for #AdamAnswers!
#AdamAnswers is the segment on the show where I answer your questions about medicine, no matter how obscure, trivial, or profound. I had a lot of great ones submitted this month — and don’t worry, I plan to get to them all. But I wanted to address one from Dr. Chris Chiu, who asks, “any cool things that we can learn about Hickam’s dictum?” And I love this, since it gives me another chance to talk about epistemology, in particular about philosophical razors and anti-razors. And I have to give extra points to the field of philosophy for having awesomely tortured metaphors. I imagine you would go to an anti-barber, who would shave you with an anti-razor, leading to an awesome full beard?
So Dr. Chiu, to talk about Hickam’s dictum, we first need to talk about Occam’s razor — one of the most utilized, and I would argued misunderstood, tools that we have. The man who invented, or rather popularized, Occam’s razor is usually just glossed over, but that’s unfortunate since he’s fascinating guy. William was born in 1285 in Ockham, England, not a noble, so therefore with no last name. A brilliant young man, he joined the Franciscan order and went to Oxford where he studied theology. At Oxford, he specialized in logic and was a fierce debater, using what he called his Law of Parsimony. Though the famous Latin phrase came later, his Law of Parsimony can best be described as: Entia non sunt multiplicanda sine necessitate, or entities are not to be multiplied without necessity. So let’s unpack this for a second — an “entity” is a hypothetical explanation for something. So what William means is that if you have two equal explanations, and one involves a hypothetical construct, or entity, and the second does not, the second is to be preferred. The only “necessary entity” to William was God. William used his Law of Parsimony liberally — first up were Platonic forms, which William argued were unnecessary entities. Many theological arguments of his day also included increasingly hypothetical arguments; these were on his chopping block as well. He even wrote a commentary that challenged natural structure as described by Aristotle.
At this point it is 1319. England is still Catholic, and I’m sure you can imagine what happened next. William is expelled from Oxford, and brought to Avignon for an audience with the Pope for his heretical views. Seeing what’s coming, he flees, and is excommunicated. Again, this is firmly in the medieval era; the protestant Reformation is still centuries away. But William sought an audience with the Holy Roman Emperor Louis IV, and reportedly plead: “Emperor, defend me with your sword, and I will defend you with my pen.” Louis was apparently swayed by this, and William continued to work in Munich until his death in 1349.
The Law of Parsimony continued to be influential in philosophical and theological debates over the ensuing centuries. In 1852, realizing that William had in many ways divined scientific thinking, the historian William Hamilton renamed the Law of Parsimony to “Occam’s Razor,” and the name has stuck ever since, even though the use of Occam as his name is similar to calling Leonardo “Da Vinci,” an anachronism, grating, but it’s still more catchy than “William’s Razor” (or, for that matter, a Dan Brown thriller called “The Leonardo Code”).
Occam’s Razor has been used and referenced in a variety of fields ever since. Einstein used it to justify special relativity, and he was famously paraphrased as saying, “everything should be as simple as possible, but not simpler.” “Penal parsimony” was used to argue for just prison sentences. And despite William using his Razor to justify the existence of God, it’s been used to argue for atheism. Mathematicians have also redefined Occam’s razor in terms of Bayesian probability — rather than just a helpful mental heuristic, it’s mathematically true. Finally, for our purposes, Occam’s Razor is frequently used in medicine.
I unfortunately haven’t been able to find out just when Occam’s razor started to be used in medicine, but it’s clearly before the early 20th century, since doctors writing then reference it as something that their audience would know. Unsurprisingly, William Osler has been credited with its introduction in medicine — and while he used it, just color me skeptical as to him being the first. There’s a whole cottage industry of attributing everything to Osler. As far as I can tell, there are two major formulations of Occam’s razor that are used on the wards.
The most famous comes from Theodore Woodward, a lecturer at the University of Maryland Medical School, who famously quipped. “When you hear hoofbeats, think horses, not zebras.” And note that this does NOT come from House of God — Samuel Shem was just repeating what was already a common saying. Now mind you, I worked in Botswana, in which there are plenty of zebras and not a lot of horses, so your mileage may vary, especially if you’re in southern Africa. But in this formulation, given two explanatory diagnoses, the more common one is more likely. I usually prefer “Common things are common.” Woodward’s saying has become so prevalent that “Zebra” is used as a synonym for a rare diagnosis. Examples of this abound. If your patient comes in with a nosebleed, it’s appropriate to think about trauma before diagnosing Osler-Weber-Rendu syndrome.
The second formulation is called “diagnostic parsimony.” Basically, if you have a constellation of symptoms that can be explained by a single diagnosis, rather than multiple, the single diagnosis is to be preferred. Take, for an example, a young man who presents with acute onset of rhinorrhea, shaking fevers and chills, a severe headache, and an incredibly rigid and painful neck. Yes, it it possible that he has influenza and torticollis — but isn’t it more likely that he has meningitis?
That’s enough on the razor. Let’s get to anti-razors. The most famous comes from Walter Chatton, who was William’s contemporary and sparring partner. He developed the Chatton Principle, generally called Chatton’s Anti-Razor:
“Whenever an affirmative proposition is apt to be verified for actually existing things, if two things, howsoever they are present according to arrangement and duration, cannot suffice for the verification of the proposition while another thing is lacking, then one must posit that other thing.”
There are a ton of different anti-razors that have been developed over the centuries, and complementary razors as well — some of my favorites are Crabtree’s bludgeon, attributed in a Borgian way to a fictitious philosopher invented in 1954: “No set of mutually inconsistent observations can exist for which some human intellect cannot conceive a coherent explanation, however complicated,” and Adler’s razor, aka Newton’s flaming laser sword: “What cannot be settled by experiment is not worth debating.” But to finally get to Dr. Chiu’s question, in medicine the most famous anti-razor is “Hickam’s Dictum”: “A man can have as many diseases as he damn well pleases.” This has been attributed to John Hickam, but it actually appears to be an apocryphal saying formulated to sound like “Occam”. The other famous real example is the so-called “Saint’s Triad”, by South African doctor CFM Saint — hiatal hernia, gallbladder disease, and diverticulosis. He noted this is a series of patients, but was unable to find any unifying underlying diagnosis. His point in his lectures was that the very existence of Saint’s Triad argues against diagnostic parsimony; sometimes the patient has more than one thing. There are actually three compelling arguments that Hickam’s dictum is becoming more relevant as time goes on — the first, as the population ages, it comes more and more likely that individual patients will “accrue” diagnoses. In other words, a 60 year-old with severe abdominal pain and burning on urination is much more likely to have diverticulitis AND a urinary tract infection than a 30 year-old with the same symptoms. The second is increased immunosuppression, classically seen in AIDS, but now also seen on patient’s on immunomodulatory treatment; with the immune system turned down, it is fully possible for patient’s to have any number of even infectious diagnoses. And finally, improved imaging leads to us making far more diagnoses.
So what is it, Occam’s razor or Hickam’s dictum? If you do a Pubmed search, you’ll find a lot of case reports essentially arguing that the two are in opposition to one another. That’s not the approach that I would take. Occam’s Razor and Hickam’s dictum are both heuristics, tools that help us take care of our patients; there are situations where we should approach one and not the other, and there are situations where we should take a stepwise approach — consider Occam, and if that fails, consider Hickam. Take an AIDS patient, for example. If they present with altered mental status, fevers, and a horrible cough, I would think both PCP pneumonia and possibly an intracranial lesion. The same presentation in a 60 year-old without AIDS? My first thought would probably be a toxic encephalopathy from a pneumonia.
So thank you so much for the excellent question Dr. Chiu! I enjoyed working on this so much that I’ll be publishing a Tweetorial on Occam’s Razor and Hickam’s dictum; so check it out on my Twitter feed. And for those of you who don’t know, Chris is a producer for The Curbsiders, basically the “podcast of record” for internal medicine. If you haven’t listened, you should check it out.
That’s really it for the show! Keep tuned in if you want to hear the rest of Dr. Herzig and my conversation. OK, let’s talk about the American College of Physicians! The American College of Physicians, or ACP, is my professional society; it’s for internal medicine physicians of all stripes in the United States. What that means for you is that if you are a member of the ACP, you can get continuing medical education, that is CME, or maintenance of certification, MOC, for listening to Bedside Rounds! You just go to www.acponline.org/bedsiderounds and take a brief quiz. And if you’re an internist, either attending or resident, or a medical student interested in the field, and NOT a member of the ACP, I strongly encourage you to join — it’s a great organization advocating for physicians and their patients that I’m honestly proud to work with. And this partnership means I’ll get to continue sharing my work on medical history, and its intersections with culture and society, for an even larger group. Because ultimately, Bedside Rounds is more than just cool stories — it’s about the process by which we’re all making modern medicine, a process that we’re all in the middle of. By looking at our past, we can better understand where we’re heading. And of course, I want to give all of my listeners a big thank, because I literally couldn’t have done this without you. It says something remarkable that so many of you tune in to hear me opine about 200 year-old medical studies month after month.
I’m simultaneously releasing an “introductory” podcast, called “Episode 0.” Even if you’re a long-time listeners, give it a listen. It’s my manifesto of why I think medical history is important. You can listen to my entire back catalogue on the website, www.bedside-rounds.org, or on Apple Podcasts, Spotify, Stitcher, or the podcast retrieval method of your preference. I’m on facebook at /BedsideRounds. And I’m a little bit addicted to Twitter; come say hi @AdamRodmanMD. I tweet about medical history and evidence-based medicine, and I’ve gotten into Tweetorials in a dangerous way. Never seen a Tweetorial? Come by; I’ve written a fun one on Occam’s Razor which I’ll be posting probably by the time you listen to this.
All of my sources are in the shownotes.
And finally, while I am actually a doctor and I don’t just play one on the internet, this podcast is intended to be purely for entertainment and informational purposes, and should not be construed as medical advice. If you have any medical concerns, please see your primary care provider.