Sunday, August 04, 2013

Getting doctors to be doctors again

kw: book reviews, medicine, doctors, advice

There is a strong parallel between medicine and religion. Even more, the contrast between "doctoring" and genuine medical practice is like that between religion and faith. My favorite proverb about religion is
The core of religion is a checklist. You can hang it on the wall. A robot could perform the checklist perfectly. You cannot.
This is why I say that my faith is not a religion. Faith is a divine relationship; religion is a specified practice. What about doctoring? Doctors today make explicit use of checklists, which they often call "pathways", to "work up" a patient who comes to them with a particular "chief complaint". A wry joke going around is, to get seen quickly in an emergency room, when you tell the receptionist why you are there, make sure that you mention your chest hurts. Of course, that means you'll be there at least 24 hours and spend an extra $10,000.

If you actually try this, you will find yourself on the "Chest Pain Pathway", which involves a series of tests, including a CT scan, and an overnight stay for "observation". If you don't actually have any chest pain, you will still have a hell of a time getting off the pathway. There are hundreds (thousands?) of such "pathways".

The modern doctor will listen to you only for a few seconds, until you say something that triggers one of the pathways. If 10-15 seconds pass and no pathway is yet identified, the doctor will interrupt you with a series of questions, always of the yes/no variety. Once you mention the "right words" that trigger a pathway, even in passing, more yes/no questions follow, and pretty soon the doctor will stand and say "we'll do a couple of tests" while heading for the door. Less than a minute will have passed. If you happen to be a rambling storyteller, particularly if you are under great stress and confusion, you may not yet have talked about the real reason you are there! But now you are on a pathway, and you may never get off it (because you may die, or be sent home with "no finding").

It is estimated that at least 100,000 people die in U.S. hospitals due to "misdiagnosis". In most cases, it would be more accurate to say they died without a diagnosis, because they were killed by the hunt for one.

To literally save your life, you first have to do everything you can to stay off one of these pathways, because here's the dirty secret. A pathway is designed to minimize the chances you will sue the doctor, or at least make it very unlikely that you could win a malpractice lawsuit. They are NOT designed to lead to a diagnosis. But, it is becoming apparent that pathways themselves are a kind of malpractice.

Drs. Leana Wen and Joshua Kosowsky call this "cookbook medicine". I call it "robot medicine". A robot can do it; you don't need a doctor for that (look up "Symptom checker" and try the one at WebMD, for example). Robot medicine is a practice of "ruling out", and you'll hear a doctor say, "This may be a heart attack," (you mentioned chest pain) or, "We have to rule out a cranial hemorrhage" (you used the words "worst headache"). Drs. Wen and Kosowsky have written When Doctors Don't Listen: How to Avoid Misdiagnoses and Unnecessary Tests to help you and me learn to stay off such pathways and drive a doctor to diagnosis.

Another proverb you probably know: To a man with a hammer, everything looks like a nail. The very availability of CT scanners, and the readiness of insurance companies to pay for CT scans, means that scanners are massively overused. Ditto for MRI and PET and other scanners, and for increasingly specific (and costly) blood tests. Excess CT scans are particularly worrisome, because even the most modern, low-dose scan exposes you to 500-1,000 times the radiation of an old-fashioned X-ray. ER's ought to have a chart on the wall:
  • Got chest pain? Prepare for an overnight stay, 3-4 blood draws and a CT scan. Equivalent radiation dose (ERD): 400 chest X-rays.
  • Terrible headache? 2 blood draws and a CT scan of the head. ERD 500 dental X-rays.
  • Abdominal pain with diarrhea? Prepare to spend the night, plus 3 blood draws, a sigmoidoscopy [actually a good idea!] and an abdominal CT scan. ERD 600 X-rays.
And so forth. Just for grins, if the radiologist has any trouble reading the CT scan, or notices any anomaly, you'll get a second CT scan (or if you are lucky an MRI scan: there is no radiation), with a different IV of contrast "stuff". Many folks are allergic to the contrast injection, such as my wife, who is allergic to both media that are used for kidney X-rays. Meaning she can never have another X-ray looking for kidney problems.

Properly used, CT scans save lives. But they come with a known risk: a 1-2% increase in the chance of getting cancer after 20 years or so. If, as some contend, 98-99% of CT scans are unnecessary, then it is a wash, a life lost to cancer for every life saved.

In the past it was not so. An early chapter of the book traces medical history over the past 20 centuries. Of the four stages of medical practice, the third may have been the best. From the time of Pasteur and Koch, and into the early era of antibiotics, doctors were very diligent to get a good "history" from a patient, plus they had a growing number of effective remedies, both medicines and surgical techniques. Prior to the 1960's, diagnosis was king. The most effective doctors had a quality known as Augenblick (German for "eye blink"). My uncle's father was one such. He could diagnose many, many conditions at a glance, yet he always took time to hear what the patient said (or the person who brought the patient in, if the patient was unconscious). The history might modify the eyeblink diagnosis. He took nothing for granted, but also had enormous common sense. If a patient arrived at the hospital with a note from him, the doctors there knew the diagnosis was a good one.

Diagnosis is still king, but the king is in hiding. Doctors practice defensive medicine, particularly in the ER. This consists not in driving to a diagnosis, but in "ruling out" rare conditions that could lead to a lawsuit if they are missed. The way a good entry interview ought to go is simple. You tell your story and the doctor listens to all of it. If you aren't a good storyteller (some doctors use the disparaging term "poor historian", but such doctors ought to go be plumbers or something), the doctor will ask questions to move the story along; preferably, questions not of the yes/no variety. At some point, the doctor will have a short list of possible diagnoses. The questions that follow are designed to distinguish among these competing diagnoses. This is called differential diagnosis. At some point, one idea is the most likely, and this is the working diagnosis (sometimes there will still be two, and that may be the actual case, that you have two things wrong at once). A good doctor won't just do all this in his head, but will discuss the possibilities with you, and will tell you the working diagnosis, or perhaps the last two possible diagnoses, with suggestions for making a final distinction. Only then can effective treatment begin.

It helps to learn to be a better storyteller. Consider chest pain. It is sure to trigger a "chest pain pathway", unless you first mention that you helped your cousin move yesterday, and carried one end of her piano. If you can also press a finger to one spot, saying, "It hurts the most right there", you have a chance of knocking your doctor off the pathway, and getting appropriate help for a pulled muscle in your chest. If not, you could be swept along with the momentum of the pathway, and go home a day or two later, with a CT scan or two under your belt and at least 3 fresh needle sticks. And, nothing for your pulled muscle.

The word "doctor" means "teacher". Prior to "modern medicine", a doctor taught all her patients how to recognize what symptoms meant, and what to do about it. Doctors didn't ask yes/no questions until late in the process, to distinguish provisional diagnoses from one another. During the history-taking phrase, the questions were open-ended, intended to get every scrap of information you could deliver.

To help your doctor arrive at the right diagnosis, and thus to treat you for the right condition, these days you have to help out, a lot! To this end, Drs. Wen and Kosowsky spend half the book training us to use their "8 Pillars to Better Diagnosis":
  1. Tell your whole story. This means making sure the doctor gets it all, particularly that which worries you the most (don't let embarrassment hold you back; tell it all). And don't answer yes or no to yes/no questions; use them as a springboard to tell more of your story. When the doctor interrupts, interrupt right back (as nicely as you can).
  2. Assert yourself into the doctor's thought process. I'd have used the word "insert", but the authors are emphasizing that you have to be assertive, almost pushy, to keep the doctor collaborating with you.
  3. Participate in the physical exam. Of course, show the doctor "where it hurts", so to speak, but also ask what he is thinking. Ask for jargon words to be explained.
  4. Make the differential diagnosis together. Ask for the doctor's list of ideas. If none make sense, say so and ask, "What else could this be?" The initial net needs to be broad, but seek a balance. 100 vague possibilities are not a differential diagnosis.
  5. Partner for the decision making process. Continue to ask how this or that possibility can be most simply eliminated, or how any one could be confirmed.
  6. Apply tests rationally. Don't agree to "ruling out" tests for low-probability conditions unless your doctor can make a good case why every symptom you have described fits that condition, and none of them in themselves make it less likely. Be sure you understand what every test is meant to distinguish or determine.
  7. Use common sense to confirm a working diagnosis. The visit is not over without a working diagnosis. But it has to make sense.
  8. Integrate the diagnosis into the healing process. The doctor must tell you what to expect from the treatment she prescribes or recommends, and how it relates to the diagnosis. The working diagnosis may not have been the right one, so knowing what to expect can clue you in that something isn't quite right.
Nobody will remember these points during an ER visit, or even an office visit, so one third of the book consists of exercises and practices to help you get very familiar with the process. No longer can you and I afford to passively "let the doctor do the doctoring", because we will almost always find ourselves being pushed toward one of the "pathways". These pathways are not helpful to you, but to the doctor's malpractice insurance carrier. Throughout, the authors stress that we must be respectful with the doctor, not confrontational, because, after all, we went to him for his help, and such help will be harder to elicit if we offend him. If your doctor is thin-skinned and arrogant, do everything you can to change doctors.

My most important take-away from this book is a question that I need to ask: "Doctor, what will this test determine? What is the diagnosis you wish it to confirm?" If the doctor intends to "rule out" something, don't agree to it. Demand (nicely) a differential diagnosis, and ask where the test fits into that.

I am in the process of getting a new doctor. A couple weeks ago my wife made an appointment for me (shoulder strain) with my primary care physician. John is a good doctor, one who listens much better than most, and we've developed a good working relationship. But a week later, his office called to say that he had retired, there is now a new doctor taking over the practice, and did I still want the appointment? Well, what choice do I have? Now that I am on Medicare, it is hard to find a doctor who will take a new patient. I accepted the appointment. Now I have a new doctor to train. I'll make copies of the relevant parts of WDDL and study up, so when I see him a couple days from now, I can start off on a good footing and gain an appropriate collaborative relationship.

No comments: