Too Important to Measure

“So, please join the movement to ban productivity from medicine. We are not producing anything. We are caring for patients who need our full attention.” This is the concluding appeal from “It’s Time to Ban ‘Productivity’ in Medicine,” by Robert Centor, M.D.

This type of rhetoric is designed to imply that anyone who disagrees hates people – or, in this case, patients. Often it is children or mothers or women in general. But the basic formula is a false dichotomy, where if you do things my way, you care for people; if you do anything else, all you care about is getting rich while harming others.

Profit is not the enemy; at least not any and all profit. “Profit” at its simplest is just a way of saying thanks that is tangible: the doctor gets paid for the health care.

Businesses today can be so multi-layered and convoluted that it’s easy to imagine the “profit” only means that some man in a suit somewhere far away from the real work gets to buy a new yacht. While this can be one of the outcomes, and we can agree for the sake of the discussion that it’s unfair, it’s not the simplest or most basic meaning of profit. In the simplest form, it means that the guy in the suit can pay the doctor for his work and not lose money. This is good, because if the guy in the suit loses money and expects to keep losing money he quits paying the doctor.

Then the doctor would need to get his or her money directly from you. And the doctor would need to be really sure that he or she is getting enough money to pay off the medical school debt, and also make all those grueling years of school worthwhile. It’s hard to be sure of that unless you get that guy in the suit involved again. Paying for the doctor’s schooling just makes this harder to talk about but doesn’t really change the facts: whoever teaches the doctor needs money, so if the patient who pays taxes to pay for the schooling to be provided for free to the doctor and pays more taxes to pay the doctor so that the actual doctoring is free – it’s just more complicated, but at the end, if the whole thing is not making a profit, it is losing money and will have to shut down.

There are big, hard to manage problems in health care in the USA today. And it’s a lot harder to measure whether doctors are doing a good job than it is to measure whether widget-builders are building widgets.

Those two truths do not justify this self-entitled appeal to “ban” productivity measurement in health care.

The complaints that Dr. Robert Centor has are basically the same complaints that everybody everywhere in every job has when their job gets measured. Measurement always introduces the possibility of perverse incentives. When you measure how many widgets the widget-builders make, they have a tendency to make more; when you emphasize the measurement of how fast they can make widgets, they make more at the expense of product quality, safety, and improvement to the product itself and the production process. Improper measurements of “productivity” were one of the major problems in American manufacturing that were called out by Toyota (and other Japanese manufacturers) when they overtook American car manufacturers in product quality and reliability.

Thus, the Toyota Production System never recommends measuring how many units are produced as a good measure by itself. At the most basic level, the Toyota Production System admonishes you to understand how many are actually wanted and not to build too many. Along with that, if the widgets you make do not work they do not count as “units produced” – quality matters. Furthermore, if you haven’t done your research to make sure the widgets do something valuable to your customers, anything you make is waste no matter how expertly you make it.

Transferred over to health care, the first thing we can observe is that measuring how many patients doctors see is probably the same as measuring how many junky unsafe cars factory workers can produce. It’s measuring the wrong thing. How much only matters if you are making the right stuff. So if we are going to learn any lessons from manufacturing productivity – any of the lessons on productivity and quality from the last five decades or more – we need to figure out how to measure whether doctors are doing the right stuff before we try to measure how much of it they do.

Measuring quality of care is a lot harder than measuring quantity of care, no doubt.

But even measuring quality of manufactured goods is harder than it sounds at first. Most of the cars sold in the world today don’t need to be able to drive any faster; a car that can go 125 mph is not worth any more than a car that can go 95 mph to most people. But a car seat that is comfortable for hours at a time is probably worth paying for. That’s harder to figure out, but it matters more for the real meaning of product quality (what customers are willing to pay for).

I’ve said it before but I probably can’t say it too many times. Measuring quality of care is hard. But think for a minute about the opposite. What if we don’t measure care quality at all, ever?

Well, if we don’t ever measure the quality of care, that gets us right back into pre-scientific medicine: “Sounds to me like you need to swallow some cat dung. Hope it works for you!”

Instead, although measuring care quality is hard, and we can’t do it perfectly, we need to measure care quality and keep finding new and more accurate ways to measure it.

But what about costs? Leaving aside the vast and unhelpful complications of the current health insurance system, if I go to Dr. Jones for corrective lenses and he prescribes glasses that adjust my eyesight to 20/20 and charges me $200, but (for science!) I then go to Dr. Smith and he prescribes glasses that also bring my sight to 20/20 and charges me $500, which is a better deal?

Okay, so the same work for less money is better productivity.

Tell me again why we shouldn’t measure productivity?

What we need to do is have continuing conversations about how to measure real productivity. Yes, there is definitely a possibility that measurements could lead to perverse incentives; managers need to recognize that and account for it. Managers who emphasize simple measures of “productivity” in factories will see quality go down, and DID see quality go down, and lost their business to competitors who paid more attention to quality. If you are a health care provider working for an administration that measures “productivity” only in quantity, the problem is not the measurement: it’s your management. And they will lose their business. (Provided they are allowed to lose it; which means they have to lose money if they don’t satisfy patients; which requires patients are allowed to choose providers whom they think give better care; but that problem is for another discussion.)

I was recently asked how I would apply Lean Six Sigma concepts to reduce patient visit time, if given the chance. Part of my answer was: who says patients want their visit time reduced? Most people I know want to spend more time with the caregiver and less time in the waiting room, so “reducing time” is a poorly-stated goal.

The root cause of the problem Dr. Centor is observing is not the use of measurements; it’s that the same interest group gets to determine limits on what the patient needs and how much the doctor will get paid for it. Insurance companies can limit what they will “accept” in billing, which effectively means what the doctor is allowed to do (to be paid). If patients paid directly, they would pay more to spend more time with the caregiver. Since the payers are managing by numbers, not by experience, they need countable things like tests prescribed; add that with a very natural desire to make sure you’re getting bang for the buck (which is, in this case, measureable health care per dollar of reimbursement), and presto! You now have an incentive to prescribe as many expensive treatments in a short period of time as possible.

Misuse of measurements will always be a risk, especially when money is involved. The right approach is not to stop measuring altogether, as Dr. Centor suggests, but to think carefully about what you measure and think twice as hard about what it means. Problems with measuring performance do not apply in a fundamentally unique way to health care providers; there are problems with measuring engineers, HR communications, and office workers in general; right on down through to factory workers. Managing by the numbers makes no more sense than driving your car by staring at the speedometer. You need the speedometer to keep your senses in check, but you need to stay aware of the context to know whether 45 MPH is too slow on the highway or too fast in a school zone.

In the case of Dr. Centor, his rallying cry was addressed pretty well by commentator David Pogge, who wrote in part: “However, the implied alternative appears to be a system based on the notion that every provider knows what is ‘optimal’ and their work will drift towards providing that optimal service if external pressures are removed. This idea is either self-aggrandizing or naive. Healthcare providers are human, and therefore they are as flawed, selfish, shortsighted, lazy, and prone to misjudgment as all other human beings. To think otherwise suggests that one has never actually worked with real people in a real healthcare setting.” People, whether patients or doctors or factory workers, cannot be managed well by numbers alone. But people, even doctors, need numbers to provide objective feedback. Sometimes the numbers will be misleading, and sometimes the numbers will be irrelevant; but medicine without measurement is just superstition. A doctor should know better.