Observational Detectives

Lisa Sanders, the diagnosis columnist for New York Times Magazine (and, I should disclose, my close friend), has just published a truly interesting book, Every Patient Tells A Story, on how good doctors go about making difficult diagnoses.

I personally benefited from Lisa’s own diagnostic prowess just last week, when I mentioned to her my wild nightmares at the height of my battle with swine flu (which I wrote about here). I also told her that I awoke the next morning with a terrible sprain in my hip. I felt as if I’d been through a Jacobean “struggle with God”. I figured I must have been flailing around during my dream and pulled a muscle.

Lisa’s responded, “I bet you were taking Tamiflu.” While I had diagnosed the flu as the cause of my bad dreams and by extension my sore hip, Lisa figured that Tamiflu might have been a contributing cause. She’d heard that some patients had reported very vivid dreams while on Tamiflu. Maybe both my nightmares and flailing had occurred during a Tamaflu-induced delirium. She told me that real dreams occur during REM sleep when your muscles are basically immobilized — so it is unlikely that I would have been physically flailing around and stressing my hip while I was in deep sleep.

Before reading Lisa’s book, I thought that the process of diagnosis was much more a kind of formulaic pattern recognition: if you can remember that Lemierre’s disease normally presents with pain and swelling on one side of the neck, then you are going to want to include Lemierre’s in your differential for any patient with this kind of pain. In Super Crunchers, I conjectured that predictive analytics was going to invade physicians’ autonomy in the front-end job of making diagnoses — much as number crunching has already done in the back-end job of choosing the best treatment. But Every Patient Tells a Story convinces me that I missed the important and non-formulaic role that good physicians will continue to play in acquiring and analyzing the raw information needed before any formulaic algorithms can be applied:

Medicine — to the extent that it can be called a science — is a sensual science, one in which we collect data about a patient through touch and the other senses according to a systemic method in order to make a diagnosis.

Lisa shows how touch, sight, hearing, smell, and even taste are critical tools for diagnosis:

We are trained from a very early age to avert our eyes from abnormalities. Children are fascinated by people whose appearance differs from what they’ve come to expect. And we teach them to ignore that interest. My daughter Tarpley once asked a cashier if she was a man or a woman. My husband flushed with shame for the discomfort it cause the homely, hirsute woman. … Afterward he explained to our daughter just how much that kind of comment must have hurt the woman. She doesn’t ask those kinds of questions anymore. She’s learned not to stare.

Medical school forces you to undo that training. You mustn’t avert your eyes from abnormality. You need to seek it out. You need to figure it out. And it doesn’t just turn off when you leave your office. I frequently (quietly I hope) point out to my husband pathology that I see on the street — the rolling gait of a man with an above-the-knee prosthesis; the strange gray-tone tan of a man with iron overload syndrome … the schizophrenic woman’s restless lips and mouth, a long-term side effect of many antipsychotics.

I recently had a chance to hear Lisa give a fancy lecture to Yale’s Sherlock Holmes Society. To a packed lecture hall, she argued that TV was experiencing a resurgence in what she called “observational detectives.” Shows like Lie to Me, Monk, and especially House featured sleuths who, like Sherlock Holmes, relied primarily on their powers of observation.

Lisa knows of what she speaks when she comes to Gregory House, the brilliant but emotionally dysfunctional diagnostic detective, because she’s been a consultant on the show and her New York Times column was one of the inspirations for creating a show centered on trying to figure out what was causing a patient’s illness. I’m a big fan of the show, so Lisa blew my mind when she pointed out how directly the House character was based on Sherlock Holmes: both have names that sound like a domicile (Holmes/House); both are addicted to narcotics (cocaine/Vicodin) and both are sufficiently dyspeptic that they have only one friend (Watson/Wilson). In the opening episode of the show, House encounters a man whose skin is orange and in truly Holmsian fashion not only deduces the illness but predicts that the man’s wife is having an affair (because otherwise she would have noticed his new skin tone).

In the lecture, Lisa wondered why observational detectives were coming back into vogue now. I think Jerry Seinfeld might be part of the answer. After all, it was Seinfeld (following in the footsteps of George Carlin and others) who made observational comedy so dominant. Seinfeld is the guy who spotlights the quirky things that the rest of us miss.

Indeed, in a world with so much observational detecting and observational comedy, it maybe shouldn’t come as a surprise that we are also seeing an uptick in “observational economics.” In many ways, Lisa’s fascination with abnormal medical conditions reminds me of Steve Levitt’s freakish economics. They are both observational detectives ferreting out the hidden side of everyday life.


Nice post.

My dad was a diagnostician. It's guided intuition with a willingness to change direction as facts reveal themselves. I remember him telling me about being an examiner for the certification board. Rather than dig up the rarest films and facts, he'd test them on something like diabetes and then go into oddities to see if they understood coherently or if was just a bunch of facts to them. My dad was really smart.

Brad Hicks

I'm a big fan of Lisa Sanders' columns. I have a friend who's a doctor with similar diagnostic skill; at the hospital where she works, she keeps saving patients that other doctors have written off because of her combination of observational skills, deductive reasoning, and sheer persistence -- she doesn't give up until she knows WHY the patient is displaying these symptoms. She mentioned to me, the other day, that she almost certainly has saved the hospital she works for from two multi-million-dollar wrongful death lawsuits in the last year.

She's the lowest-status doctor in the emergency room. Why? Two reasons. (1) She's not a specialist. And (2) diagnosis is not, technically speaking, a "procedure" that can be billed for.

Add this to the long, long list of things we need to fix about our healthcare system: the current incentives are driving our best diagnosticians to quit their work, leaving it to the patient, to nurse practitioners, or to doctors too rushed to pay attention and too unskilled to seek more prestigious work, to actually help us figure out what procedures we actually need.

And no, I don't know what to do about that.


Eric M. Jones

Just to be sure you understand the Holmes-House thing was intentional. House lives at 221B Baker Street.


You just blew my mind.

Chad Bergeron

Even more telling, both Holmes and House live at 221B. Both are known for being drug addicts and musicians. House is not just an echo, but an homage.

anon MD

I agree with Brad that the system is broken, but it is impossible to fix if even patients are not appreciative of our efforts.

I once diagnosed an early lung cancer by finding a lymph node on the patient's collarbone during a checkup. The tumor was so early that she required only surgery, no chemo or radiation. Instead of thanking me, she stopped seeing me because she accused me of "blackmailing" her into office visits (I told her I needed to check her blood pressure occasionally in order to refill her blood pressure meds).

Patients expect doctors to catch these little things and save their lives, but are not willing to place more value on primary care. Instead, they are eager to sue when rushed visits (as doctors try to pack more into their day to make up for declining reimbursements) lead to a missed diagnosis.

adina (Med Student)

Nice article. Algorithms may one day replace humans in determining a diagnosis, but only once a series of symptoms and physical findings have been entered in a database. For many years to come, it will still take a human being to notice, elicit, identify, and input those findings.


Another observational detective show I quite like - Psych - a bit lighter, but the principle still applies.

I had never picked up on the House / Holmes connection - probably explains why I am such a fan of House.


Isn't House the poster child for why (supposedly) health care is so expensive? Takes four or five tries until the correct diagnosis is arrived at, usually with multiple expensive diagnostic tests along the way...


My sister suffers from chronic migraines. Towards the beginning, she also suffered enormously from muscle cramps over her entire body - bad enough to keep her in bed for days or even weeks at a time (and she's always had a high pain threshold). This went on for months until she tried a new doctor, who knew that severe muscle cramps were a known (but rare) side effect of the interaction between a drug one doctor had prescribed for anxiety and another drug another doctor had prescribed for migraines (or something along those lines - I'm not sure of the details.) Her meds were adjusted and the cramps went away. Both of the prescribing doctors had known about the other medicines she was on, but apparently missed that connection.


i love the house and holmes part...enlightening..

Avi Rappoport

I think it was Lowell Thomas who said that medicine is the youngest science. Unlike physics, there are few pretty equations and determinable events. Our culture is really amazing at a lot of things, like programming and engineering, but not so much on unknown semi-random systems, like human bodies and human stock markets.


I had a conversation at a friend a month ago about this very topic.
My friend could talk on any subject with almost total recall.
I made the observation that this reminds me of both Holms and House and of coarse he made the further connection that they both came under the control of mind alterating drugs..
It was my contention that both show signs of autism or what we call now Asperger's syndrome.
You can't learn to be a observational detective.
It's an ability based on how the brain processes information (seeing connections between seemingly unrelated facts) with total recall putting those relationships in a context that enable the individual to deduce something that is not known like a medical condition or the solution to a crime.
It's my contention that the above is also a description of Asperger Syndrome which I made the observation might be true with my friend and me.
This was confirmed when he told me that he has relatives who are autistic which I did not know but didn't surprise me.
Back to the subject.
Both Holms and House would be unable to do what they do so well were they not Autistic . .
By the way there is an individual in history who also displayed these very characteristics and that was Einstein and we know what he deduced.


Jean Gogolin

Re Avi Rappaport's comment: It was Lewis Thomas, a physician, who called medicine "The Youngest Science" and wrote a book with that title. I have it on my office bookshelf and it's a great read. I am -- ahem -- old enough to remember Lowell Thomas, a longtime radio journalist.


Last time "observational mysteries" were this big on TV, a cigar-chomping half-asleep detective in a trench coat was saying, "Excuse me, just one more thing..."

a physician

Most patients, in fact many people at cocktail parties, who have "only a quick medical question", don't realize that taking the time to ask, learn, and notice, let alone synthesize it all and communicate an appropriate management plan for an individual patient, even when there is not a challenging diagnosis that could miss something crucial, is what provides the best medical care. Our health care system has pretty much eliminated that option with exponentially less compensation for thinking than doing procedures, necessary or not.

I'm afraid that House is the poster child for why it can be so difficult to find a good doctor. First, there is no compensation for the hours of time he and his team spend trying to diagnose and help a patient. Second, his behaviour towards patients would've had him fired in the real word long ago since most patients and all employers care more about "satisfaction" surveys than actual medical competence. Those seem to be the reasons why most physicians I know think that House is the most unrealistic doctor show they have ever seen.

-Another doctor



"...why observational detectives were coming back into vogue now..."

I think for the the same reason movie characters like Batman and Superman exist: they have superpowers. The ability to observe and find the little telling details may be a gift or an acquired skill or both but in any case very few people have it. So to most of us who don't have this ability, it looks like a superpower ot magic. And people are fascinated by magic ;-)

Lisa Sanders also raises an interesting point about the parents teaching their children to ignore the abnormalities. I am wondering: would more people have the skills to observe if the society did not force us to avert our eyes from the different?


Look into, "The Medical Detectives" - http://www.amazon.com/Medical-Detectives-Plume-Berton-Roueche/dp/0452265886/ref=sr_1_1?ie=UTF8&s=books&qid=1250790366&sr=1-1


I was watching a medical drama with a college friend several several years ago. Homeless men were getting sick at a homeless kitchen for breakfast. I said to my friend, "I read [in The Medical Detectives] about a similar case were the wrong substance was was placed in the salt shakers by accident, and some of the men put 'salt' on their oatmeal, and that was the mystery that was making them sick"

Unfortunately, I blew the ending, since the drama used that very story for their plot.


One of the things I've enjoy seeing is people's interactions with alternative doctors. I expect at least part of the reason for the growth of alternative medicine is the more detailed questions and discussion between patient and doctor.

We don't just want to know it hurts, I want to know whether the pain is pinching, stabbing, intermittent. Whether it gets better or worse with temperature changes, cold, hot, time of day. Whether it moves or is stationary, etc. All those things together provide a much better diagnosis than the if it hurts, let's get an MRI diagnostic model.

Dr Zaglossus

The diagnostician as Sherlock Holmes has been a time honored analogy. (And the House - Holmes thing is definitely old. By the way, Holmes' addiction was Morphine.)

And, to date, computers have yet to out do (good) physicians in diagnosis.

There is another sense we physicians use, the gut, intutition. whatever you wish to call it. I've often thought that a good part of residency is simply to see enough patients. There is something about that initial impression - sometimes it can be analyzed, sometimes, not so clearly so. (And, yes, that initial impression can be quite off the marked based upon a variety of factors, stress and sleep deprivation being just two.) The initial eval is: About to Crash or not. And if not, how serious is this.

I really feel that the bullet-based check-offs that have developed thanks to the AMA's CPTs has really hurt the teaching of history taking and thus medicine in general. Instead, we rely on various technologies. But, that story is the key - how well can we get the information we need to almost vicarously or virtually re-enact what was going on in the body at the time.... Some patients are not only good self-observers but articulate enough to convey that without difficulty. Others, well, are not so gifted and thus it requires a gifted interviewer to bring that out. And with each retelling, a bit more detail comes out.