Sunday, January 4, 2009

Has medicine become "industrialized"? An intro to hospitalist medicine

Sorry in advance. This is going to be a long one.

In July 2007 on, a website for medical professionals, there appeared an article entitled “The Disappearing Doctors” (apologies in advance for the link, a subscription is required). Written by Jane Gilsdorf, a pediatric infectious disease specialist at the University of Michigan, the article discusses the complications that have arisen from recent changes in scheduling the work hours of residents—what is now almost universally called “the 80-hour workweek” based on the most important recommendation of one of the highest governing bodies in medicine. Mostly, the article is a lament about the current state of physician training wrought by the 80-hour workweek, where residents are often being shuffled between outpatient clinic and inpatient hospital coverage, covering for other residents who need to leave the hospital due to maxing out on hours, not fully understanding their patients’ conditions due to the cross-coverage. She argues, not without some justice, that this is not the optimal learning environment for an apprentice doc, noting:

“[Residents’] heads are crammed with the facts they've learned during medical school, but they can't see firsthand the course of a birth or a gall bladder attack or the phases of recovery from a surgical procedure and then integrate those facts into informed decision making. Instead of producing physicians with high professional standards who see their patients through to the end (of labor, of an operation, of an illness, of a life), the current system is creating a legion of shift-worker physicians who leave when the clock strikes a certain hour rather than when the job has been completed.” [my emphasis]

For what it’s worth I don’t completely buy her argument. For starters, the contention that one must see the “entire” course of a given illness, when taken to its logical extent, should mean that residents should never leave the hospital. After all, until a patient is out the door and restored to health, the course of illness is running and should require constant observation. Many hospitalizations span days; complicated ones last for weeks or months. Why should residents be allowed to follow the course of a hot gallbladder, which usually can be followed in two or three days, but not be expected to stand watch to the full storm of, say, a patient who develops Acute Respiratory Distress Syndrome in the midst of an influenza infection, and could require three weeks of touch-and-go care? (Of course, the term “residents” comes from the fact that, in the early part of the 20th century, they didn’t go home; they “resided” on the hospital grounds.) But we live in theoretically more enlightened times, so physicians-in-training gotta have some time off. The question is how much, and how do you schedule coverage in the interim.

But most of my issues with Dr. Gilsdorf are quibbles, and I think her points are well taken (despite having been trained more-or-less at the beginning of the implementation of the 80-hour workweek). I bring up her essay here not to discuss the merits or drawbacks of the new work-hour system for residents (though I promise to do this down the road, as I think few people outside of even residency programs really understand the implications of the changes). I wanted to use her essay as an introduction because of the last portion of her statement above, where she worries that “the current system is creating a legion of shift-worker physicians.”

By my reckoning, her worries are somewhat out of date. The current system—and by that, I mean the medical system well beyond residency programs—has already created a legion of shift-worker physicians. The old system, which I will briefly sketch out below, is largely becoming a vestige of the past. We’re already well into the new age of medicine, at least in the hospital, where the sickest patients reside.

Michael Pollan’s influential book The Omnivore’s Dilemma was not the first but has become one of the most famous books to turn a critical eye on the system of food production and distribution in the United States. Those who have read Pollan will recognize a meme throughout his writing that can be summarized as: food has become highly industrialized, just like cars, televisions, trash collection, energy production, you name it. Pollan’s analysis is that such industrialization of food is, on the whole, bad for the country, and bad for its people. It’s a remarkably provocative argument, one which I won’t try to evaluate here.

The question that Pollan’s book raises with respect to my own profession, in terms of Dr. Gilsdorf’s observation that medicine is turning into shiftwork, is this: has medicine become industrialized? And is that, perhaps like industrialized food, bad for people?

It’s a big question and I’m not going to try to take on this whole issue in one gulp. But I will try to explain to people what is becoming the model of inpatient medical care. For the lay audience out there: this is more-or-less what you’re going to experience as a patient if you get admitted to a hospital today. Many hospitals these days have turned to a shiftwork model, similar to the kind of system that Dr. Gilsdorf decries. It falls under a new subspecialty called “hospitalist” medicine. While that term was only coined just over ten years ago, it’s now becoming the standard in medium-to-large hospitals throughout the country, and my guess is that within the next generation it will almost completely replace the old model of inpatient care.

Keep in mind, as I give an intro to this, that I am one of these hospitalist-shiftworkers, at least on a part-time basis. I am one of the cogs in this maybe-industrialized arrangement.

First, what was the “old” system? Let’s say it’s 1970 and you have some bad belly pain. Assuming you have a primary-care doctor (what in those days would have been called a “GP,” for general practitioner, or just “internist”), you call up the doctor’s office and explain what’s going on. At that point, you may be asked to come to the office for an evaluation, or sent straight to the hospital and admitted directly “under” that doctor’s care. The doc will call the admitting office, arrange for the bed, speak with the nurses and secretaries on the floor and “give orders”—the top-down command model in full force in those days—for what the doc wants done. You’ll get one or a few x-rays, maybe a surgeon will drop by to see you, you’ll have some blood drawn. There weren’t cat-scans or MRIs back then so you will spend much of the time lying in bed (unless of course you go to surgery for that hot appendix). Every morning your outpatient GP will drop by to see you, examine you, talk about your improvement or lack thereof, and then the GP will sit down and write a brief progress note and any new orders for the day. The only time you will ever see a physician that isn’t your GP is if a consultant, like that surgeon, is called in. In other words, your “outpatient” doctor is your “inpatient” doctor—there wasn’t even a distinction in those days. When you’re better and it’s time to go home, you schedule a follow-up with Doc GP and when you are seen, the doc knows exactly what went on during the admission, because the doc was there for the entire course.

Now fast-forward to 2009. You have another bout of belly pain—get that fixed, will you?—and call your internist. As before, you might get into the office for an outpatient eval. You also might be sent to the hospital directly and admitted under that old system. But what is also just as likely is that you may be admitted to a hospitalist, someone you’ve never met before, who hasn’t been your doc for 40 years, who is going to be responsible for your care. Doc Hospitalist takes care of strictly inpatients for his or her job, full-time. Depending on the hospital and the vagaries of the season, that doc may have as few as a handful or as many as 20-plus patients under his or her care. Throughout your hospitalization now, you are under the care of the hospitalist.

Okay—so far it doesn’t sound so industrialized, does it? You’re just replacing one doc for another. And it’s helpful to understand just why this is the new arrangement, due to some changes in medicine as well as shifting demographics in the US in the last 30 years. (This isn’t meant to be a definitive explanation but just give the broad brushstrokes.) First, medicine is a lot more technologically complicated than it was back in 1970, with many more procedures to be performed and analyzed. We do have cat-scans, MRIs, as well as let’s-look-with-a-camera procedures (colonoscopy, upper endoscopy, bronchoscopy), guided needle biopsies, nuclear medicine imaging, and other procedures which require much more intensive management than in days of yore. A ten-minute visit in the morning and daily orders typically won’t cut it anymore. You need to have boots on the ground to manage inpatients today, with a specialist’s knowledge of which tests to order, which consultants to seek, and how to “stack” all these procedures into the most efficient process—remember, as a consequence of managed care, hospitals and insurance companies want patients out the door in the minimum time possible. That’s very, very hard to manage from the distance of the outpatient clinic via telephone.

Second, unless you are one of the fortunate physicians whose office is attached to the hospital and you can just stop by the hospital to check in on a patient during some down time in the hospital, you’re at your outpatient office in a time when more and more people live in more and more densely-populated suburban areas, with plenty of traffic to slow you down and ruin your efficiency as you drive from office to hospital and back. And I haven’t even mentioned that primary-care docs may have patients at more than one hospital.

For these reasons, you can see why the old system is being replaced, especially in large, metropolitan areas, by hospitalist medicine. So where’s the “industrialization”?

Let’s start with familiarity: most hospitalists don’t “know” their patients, are meeting them for the first time, and often will never see them again. (This is not always true of the kind of patients we somewhat churlishly refer to as “frequent flyers”—patients who often cycle in-and-out of the hospital, usually ill nursing home patients or people with chronic medical problems—but that’s typically a minority of patients on a hospitalist census.) As a rule, as long as one is vigilant, it’s not terribly difficult to learn the critically relevant details about a patient’s medical history provided good documentation and knowledgeable patients, which is not always but often possible. So while the relative anonymity of the new system isn’t directly a problem in terms of medical care, subtle problems can creep in at the margins, problems which well-informed primary care physicians who thoroughly know their patients would avoid. Bad family dynamics? The patient’s tendency to minimize his alcohol consumption? The problems created by the hospitalization of a woman who is the primary caregiver for her son, who has severe cerebral palsy and developmental delay? These are the kinds of details that don’t make it onto the “Past Medical History” list, but can be as critical to the care of the patient as the history, and are often of much greater importance to patients and their families than whether or not the patient should be on Zocor.

Another issue is one that simply bedevils the system, and it’s the same problem I alluded to in discussing the residency work-hour problem: sooner or later the doc’s gotta get some time off. That’s to be expected, but it sets up another problem with inpatient care that’s at the heart of the description of medicine as potentially being “industrialized”: shiftwork. Some call it “musical doctors” and it works like this: patient X is admitted to the hospital on a Saturday morning. They were admitted from the ER at 6:30 a.m., and the word didn’t reach the hospitalists until 7:15, by which time there is a new hospitalist on shift, only when the hospitalist tries to discuss the case with the ER doc who did the initial eval, that doc can’t be found because her ER shift ended at seven. Not so big a deal, the hospitalist can re-create the details just by evaluating the patient themselves (sometimes difficult with demented patients sent by nursing homes, but no room to go on that tangent today; demented patients and their care need their own, and perhaps several, entries). During the day, that hospitalist is the doc of record. The evening comes and the hospitalist “hands off” the service of patients to the nighttime hospitalist. Typically the nighttime doc just handles what I refer to as “tylenol calls”—generally routine matters so that patients can have pain meds, sleeping pills, and the like—as well as true emergencies that occasionally crop up. Tomorrow’s Sunday, which is “change day” at the hospital so the hospitalists can have a few days off, so there’s a new hospitalist taking over the service. Mind you, this doc has just picked up, say, fifteen new patients; he’s not just meeting patient X for the first time, he’s meeting patients A through N and patient X for the first time. (Suppose ten of these patients have a fairly long list of medical problems and medications, and you can see how much of a challenge change day is to the average hospitalist, to say nothing of the challenge it poses for the patient.)

So, within the course of 36 hours, patient X—who three or four decades ago would generally be followed by just one physician (this being the weekend it’s either the GP or the cross-covering physician, but it’s still just one)—has had at least four different people in charge of X’s medical care, with a possible breakdown in communication on the initial handoff. You can now see how there’s something approaching a factory-like quality in the system.

As I said, I work for this system, and I would like to think I provide good medical care. I’m not so much trying to expose it as an evil as I am trying to illustrate some of its drawbacks. There are very good reasons why we’ve converted to the new system. I may have scared you with the above paragraphs but the alternative scenario, where your primary care physician, who is distracted by a thousand details at the office and has only five or ten minutes to observe you on a daily basis, perhaps ordering tests that weren’t even around when he was doing full-time hospital medicine as a resident, is fraught with just as many pitfalls. But they are different pitfalls. I’m going to try to explore some of these pitfalls as this blog progresses.

Some follow-ups and then I’m done. Anyone who has gotten this far is either very interested in the subject matter or humoring the hell out of me. At the moment, and as you can see, this is structured more like a “running essay” than the kind of short, link-filled quips seen in many other blogs (the most logical extreme of which is Mickey Kaus’s blog on Slate, which I find distasteful on a variety of levels, not least of which is aesthetic). Let me know if it’s working or needs to be culled. And I do promise to discuss the commercial aspects of medicine, which I’ve only touched on in the last two entries.


  1. I agree with the issues with hospitalist, shift-style coverage. Also agree that the "old" system, with PCP managing it all from the office is essentially untenable these days with the complexity of inpatient care. Definitely agree that medicine is becoming commoditized. Answers? Anyone? Nice work Billy. - PC

  2. Your essay was nicely entertaining in style. I have been an internal medicine hospitalist for my entire medical career of 13 years. One point I would like to make is that most people would agree that a factory built Toyota will likely be more reliable than a handbuilt Ferrari, due to the quality improvements which can be identified and implemented when an act is repeated many many times. Every paradigm has it advantages and disadvantages and one of the advantages of being a hospitalist is that I take care of people with the same diagnosis over and over. Perhaps a bit boring for me, but good for the patient. One example that comes to mind is the hospital in Canada which specializes in herniorrhaphy and which claims extremely high success rates and comparably low complication rates. The manufacturing and agricultural segments of our lives have been industrialized for the very good reason that industrialization results in lower cost and better quality on average. I agree with you that the challenge now is to identify the pitfalls of this still new-ish system so that we get Toyotas and not Fiats. To that end, we need to make health care even more like a factory. People individually make errors that systems can help avoid. Caremaps, computer programs to identify drug interactions, computerized order entry, transportable (seamless) electronic health records and strong patient education actions in the form of nurses and PA's and NP's providing patient and family information about disease process, diagnosis and treatment will help us make this an industry that works well.