Press "Enter" to skip to content

Bad Data = Failing Grade

To do the math right, you have to get your numerator and denominator correct, or else things just don’t work.

Recently, in the midst of a large patient safety and quality improvement project trying to bridge gaps in breast cancer screening among our patients, we discovered that one of the large databases of patients who were attributed to us contained innumerable patients who had never been to our practice.

Officials at our institution had approached one of our major area insurers to get some data, and they were given a file showing all the patients who were “ours,” and of those patients, all of those for whom the insurer had no evidence of a mammogram being performed during the past 2 years. Based on this math, they’d given us a failing grade on breast cancer screening for this population of our patients.

But were they really our patients?

Faulty Data

It turns out, that for a large number of them, we had not had the opportunity to provide care for them, and by extension, we really shouldn’t be held responsible for whether or not they’d received age-appropriate cancer screening tests.

Somewhere in the insurer’s database were a large number of patients who’d signed up for their plan, selected a provider, but never actually made it in for any care. In many cases, by digging deeper, we were able to discover that a lot of these patients were clearly receiving care at other practices, from other providers.

Not our patients.

Similarly, for this insurer in particular, we’ve discovered that for many years, patients were coming to us and getting care at our practice, while in fact they were attributed to another provider elsewhere in the city. The repercussions of this were primarily financial, in that the insurer was, way down the electronic line, denying payment for their care, since they were seeing someone “other than their primary care provider” — despite the fact that we were, in truth, their primary care provider.

Probably, at some other practice somewhere in the city, those providers are being dinged by the insurer for not performing breast cancer screening, colon cancer screening, updating their vaccines, checking for microalbuminuria if they are diabetic, putting them on a statin if they have high cholesterol, and all the other measures we are measured by.

Many Questions Raised

One of my questions is, how did we let this happen? How do we let them tell us what the numerator and the denominator are, and who allowed us to be held accountable for this?

If the powers-that-be tell me that I’m responsible for making sure that patients who I’ve never seen have up-to-date healthcare maintenance, then they sure as shooting better give me the resources to reach out to those patients, let them know that they are “mine,” and get them into care with me. And if someone else is their primary care provider, then take them off my list.

This seems like a major flaw in all of these quality measures that the insurers are using to judge us by, and it makes me wonder how often this happens elsewhere.

They think they’ve got great data, but somehow things don’t line up quite right. We’ve all gotten many of those letters, scolding in tone, which say “Your patient with diabetes is not on an ACE inhibitor or a statin,” and then we look in the chart and discover that actually they are taking these medicines. Or they supposedly haven’t had an HbA1c test in 6 months, when in fact they had it done right here in our office lab 3 months ago.

There so many electronic systems, so many ways these data points are being collected and processed and chewed on and spit out the other end, that there’s very little chance that we are getting reliable data.

Every year, when our accountable care organization gives us our lists of attributed patients, it’s an eye-opening experience. Sure, a lot of them are names of patients that are near and dear of to us, that we’ve taken care of for years, that we see regularly. But often there’s also a lot of head-scratching, wondering how this patient got onto this list.

First of all, the data is often quite old, and I’m told it can be up to 2 or more years behind. Really, in this day and age, we can’t get up-to-the-minute data?

A recent review of one of these lists showed a large number of patients who’d never been seen in our practice, many who hadn’t been seen in several years, and quite a number who had passed away in the interim since their data had been collected.

Getting the Math Right

I’m perfectly willing to be responsible for the health of the patients I see. We are all very willing, given the right resources, to be responsible for the care of the patients we see, but isn’t what we do hard enough without also making us responsible for people we haven’t seen in many years, who we’ve never seen at all, or who have passed away?

We live in a time for healthcare where we are all beholden to these reports, and as we move toward even more pay-for-performance and pay-for-quality, this is going to begin to affect us more and more.

The contracts that are negotiated by our institutions are based on expectations of performance, and ultimately to some degree we have financial skin in the game. Even more, we want to do what’s best for all of our patients, we want to ensure that they get the best care — the right care — at the right time.

So for this particular breast cancer screening intervention project, we want to take the list of all the patients who the system tells us are ours, and find those that haven’t had appropriate screening done. We want to be able to intervene on them, to dedicate resources to them, to ensure they get what they need if they want to go ahead with this test.

But to do this, we’ve got to get the math right. We want to be able to generate reports that show which patients haven’t had the breast cancer screening field completed in our electronic health record, and set our team to work to try and help figure out whether this is correct, or whether the box just needs to be clicked.

Perhaps they had their mammogram done at an outside facility, and the report was never received in our office. Someone needs to chase it down, get the report sent over to our office, have it scanned into the system so that others can see it, and have that report satisfy the health maintenance field for breast cancer screening.

For those for whom this testing has somehow just slipped through the cracks, perhaps we need to send the patients a series of electronic reminders, an educational video, or create a focused appointment to go over the risks, benefits, and alternatives of this screening, and then have them either decide to get it done, or mark it as “declined”.

Oftentimes this screening is not appropriate, for a multitude of reasons, and that needs to be updated in the chart as well.

And then there is the patient for whom we’ve ordered it many times, and they just haven’t gone. Multiple appointments at radiology, opportunities missed, that empty order left hanging in the chart. Perhaps a really great trained patient navigator, someone with a kind and gentle touch, who could reach out to these patients, assess their barriers to getting the test done, and help guide them through to completion, if that’s what’s best for the patient.

But no matter what, we need to know our numerator and denominator, to get it right.

You do the math.