Some people are just overwhelmed by the idea of evidence-based medicine. They notice two themes when reading about EBM: 1) a lot of what we do is wrong, and 2) getting the right answer requires a lot of work. We know that “half of what physicians do is wrong,” and “less than 20 percent of what physicians do has solid research to support it.” These are astounding yet well-validated numbers. EBM promises to fix this problem, but it hasn’t and doesn’t seem like it will any time soon.

There are lots of reasons for the gap between what we know to be good practice and what is actually being practiced. Most of the reasons are just poor excuses, like the claim that docs just don’t know the guidelines so therefore they don’t follow them. These false premises might lead policy makers to focus on better information management systems, with integrated practice guidelines, and order sets for example. These types of things do help, but they don’t get at the core problem, which is the physician himself and the his attitudes.

I’m going to describe four types of physicians to help us understand the problem and how to fix it; I did not come up with this idea and I can’t for the life of me find its source after extensive searching (if you know, please email me). In any event, here are the four types:

• Type 0: this physician is a researcher who understands great depth and detail about the problem she actively researches, though she may have limited expertise and abilities in other areas.
• Type 1: this physician doesn’t necessarily do a lot of basic research, but is very astute at reading and interpreting the literature in his field, and he is able to understand the perils and promises of new research and integrate it into existing practices without making too many mistakes.
• Type 2: this physician doesn’t do research and doesn’t really read literature that much and might not even be very good at it, but she does keep abreast of the current guidelines and follows them unquestionably in delivering patient care.
• Type 3: this physician neither does research, reads literature well, nor does he keep abreast of current guidelines; rather, he mostly does what he was taught to do by others in residency and medical school and has changed his practices since then only in reaction to significant external pressures like peer-review, drug and product reps, cases with adverse outcomes, television commercials or other ad hoc and unreliable reasons.

So how are these four types of physicians distributed in the wild?

Most practicing physicians are Type 3: they haven’t changed much since their training, they don’t follow evidence-based guidelines (for a variety of reasons), they don’t actively read the literature (or if they do they pick and choose only the literature that agrees with their own practices), and they certainly aren’t doing any novel research. It is largely this group of physicians that accounts for the “half of what physicians do is wrong,” and “less than 20 percent of what physicians do has solid research to support it” segment of medicine. They drive up cost and deliver inferior outcomes. They scoff at EBM and only embrace it if it happens to support what they already believe. They are more likely to be male and more likely to be older; this also means they are more likely to be in positions of power, policy-making, and peer-review. They are more likely to be attendings in teaching programs, training a new generation of physicians to act like them.

The least common type is probably Type 0. Most good researchers are not necessarily good clinicians; it’s hard to dedicate appropriate amounts of time to both. These folks are often praised as the experts, but their expertise is usually limited to a very narrow field of interest and usually their opinions about that field are very biased because they are so personally invested in their own theories. They often cannot see the forest for the trees.

Recognizing the dangers of both extremes (Type 0 and Type 3), we have idealized the Type 1 physician, who is an expert at consuming the literature and doesn’t too easily become ensnared by its pitfalls. We want medical students and residents to become Type 1 physicians, and therefore we attempt to train them to do the hard work of reading and consuming literature and we want them to have all the necessary skills to do so: a strong background in epidemiology and statistics, strong critical thinking skills, expertise in study design and implementation, etc. But as any reader of howardisms knows, these skills require a lot of work. As it turns out, it’s just too hard to be a Type 1 physician, so very few people ever become one either because of a lack of interest or just a lack of time; this doesn’t stop a lot of people from thinking that they have become one (sometimes a little knowledge is more dangerous than no knowledge at all), but very few physicians have a skill set sufficient to critically appraise the literature.

So that leaves Type 2, right? Wrong. Because our educational system has focused on creating a generation of Type 1 physicians, but students and residents have, for the most part, rejected this, then they fall back to what they most often see modeled by their mentors and attendings: the Type 3 physician. There are barely any Type 2 physicians out there, and this is demonstrated by how poorly physicians follow guidelines.

Take, for example, labor management guidelines or pap smear guidelines or fetal monitoring guidelines in OB/GYN: most obstetricians don’t follow them (more than 80%). This is consistent with what is observed with guidelines in other specialities. We recently discussed bed rest in obstestrics, which is recommended by about 95% of obstetricians even though our guidelines recommend against the harmful practice. This is typical. Most obstetricians (most doctors) are Type 3 physicians.

One solution to this problem is to encourage students and residents to become Type 2 physicians (rather than Type 1 physicians). Being a Type 2 requires much less work and time-committment; it feels more empowering and comfortable to follow a guideline that experts agree on (rather than guessing for yourself what is correct); and it is sustainable as long as the physician has the habit of checking for new guidelines and changing her practice pattern when they emerge.

So why aren’t we doing this? The biggest challenge is the generation of physicians who are currently training young physicians. This large group of Type 3 people teach the wrong lessons time and time again. They constantly emphasize why guidelines shouldn’t be followed; they often penalize or mock residents for following guidelines; they are simply bad examples. Why do they hate guidelines so much? In most cases, they hate them because the guidelines shows them to be wrong, and there are few things as influential as the human ego.

The article I initially cited above (which is well-worth reading) relates this perspective:

The problem is that physicians don’t know what they’re doing. That is how David Eddy, MD, PhD, a healthcare economist and senior advisor for health policy and management for Kaiser Permanente, put the problem in a Business Week cover story about how much of healthcare delivery is not based on science. Plenty of proof backs up Eddy’s glib-sounding remark.

The plain fact is that many clinical decisions made by physicians appear to be arbitrary, uncertain and variable. Reams of research point to the same finding: physicians looking at the same thing will disagree with each other, or even with themselves, from 10 percent to 50 percent of the time during virtually every aspect of the medical-care process—from taking a medical history to doing a physical examination, reading a laboratory test, performing a pathological diagnosis and recommending a treatment. Physician judgment is highly variable.

That really is the problem: physicians don’t know what they are doing. Of course, they never want to appear unknowing – either to the patient or to the learner or to their egos. So they selectively fit the data around them into a narrative that supports their choices, throwing science out the window. Their students and residents learn to do the same thing, and, voila, the Type 3 is born.

So, if you’re honest enough to realize that you’re probably not going to put the work into becoming a sophisticated consumer of piles and piles of scientific journals, that’s okay! You’re normal! Focus then on just identifying and following – without question – the clinical guidelines that define your practice. You can start here at guideline.gov or check with your professional society. Need something general? Try a Google search for a disease + AAFP (e.g., otitis media + AAFP). It works.

I will point out that clinical guidelines don’t fit every patient and every presentation of diseases that we see. We have classically taught learners that guidelines are a starting point, and that the adept physician (the Type 1) will modify them as needed. Unfortunately, this idea has given license to those who would modify the guideline every time (in other words, reject the guideline completely). But this is a statistical problem. Most guidelines cover about 95% of situations (±2 SD). In those marginal cases that lie on the tails of the bell-shaped curve, there will indeed be situations where the guideline doesn’t fit. There will be patients who are in a gray area not perfectly described by the guideline. But these are the exceptions, not the rule. If you are using the guidelines in about 19 cases out of 20, you are probably a great physician. If you are not, then, well, you get the idea.