The Complexity Behind Quality Measures | Healthcare Informatics Magazine | Health IT | Information Technology Skip to content Skip to navigation

The Complexity Behind Quality Measures

September 29, 2010
by Mark Hagland
| Reprints
Jane Metzger, Principal Researcher, Emerging Practices, CSC, Reflects on a Drill-Down of The MU Quality Measures
Jane Metzger
Jane Metzger

Professionals in the Waltham, Mass.-based Emerging Practices group at the Falls Church, Va.-based CSC have recently released a new report, “Hospital Quality Reporting-The Hidden Requirements of Meaningful Use.” In the report, authors Jane Metzger, Melissa Ames, and Jared Rhoades articulate an analysis they've done of the 15 quality reporting measures for hospitals put forward in the final rule around meaningful use, released July 13, in accordance with the unfolding sets of requirements under the federal American Reinvestment and Recovery Act/Health Information Technology for Economic and Clinical Health (ARRA-HITECH) Act.

In their analysis, Metzger, Ames, and Rhoades find that “Simply meeting the Stage 1 functional requirements for data capture in meaningful use will only cover 35 percent of the unique data elements needed for the 15 required quality measures.” In fact, they note, “The biggest challenge involving the rest of the data elements is that they require physician documentation and electronic medication administration, not just in the hospital, but also in the ED.” They term these required data elements the “hidden functional requirements of Stage 1 meaningful use.”

THE ONLY WAY TO GET A REAL ANSWER TO THE QUESTION OF WHAT WILL ACTUALLY BE REQUIRED TO MEET THESE REQUIREMENTS IS TO GO TO THE MEASURE SPECIFICATIONS AND DECONSTRUCT THE QUALITY MEASURES INTO THEIR UNIQUE DATA ELEMENTS; AND THAT'S WHAT WE DID.

In fact, as Metzger and her co-authors note in their report, the titles of the “15” required measures listed in the final rule make them appear much simpler than they really are. Once one deconstructs-as they have in their report-the measures into the unique data elements described in the measure specifications, it becomes clear that many more data elements are actually required.

Even for such ostensibly straightforward elements as the provision of VTE (venous thromboembolism) prophylaxis, the reality of what is required evinces manifold greater levels of process and IT complexity that appears at the surface level. It also becomes clear that without computerized physician order entry (CPOE) beyond medication ordering, at least some physician documentation, and electronic medication administration record (eMAR) systems, the full data set will not be available in the electronic health record (EHR) as required for meaningful use. The fact that some data involve care in the ED or surgical suite/OR adds to the challenge.

Lead author Metzger spoke exclusively with HCI Editor-in-Chief Mark Hagland on the threshold of the public release of this report. Below are excerpts from that industry-exclusive interview.

Healthcare Informatics: Reading your report, one is struck by the tremendous amount of complexity lurking just below the surface when it comes to satisfying these meaningful use requirements; it's really quite astonishing.

Jane Metzger: Well, think of it this way. If you look at quality reporting, what will any hospital already have in electronic form? A fair amount. But the only way to get a real answer to the question of what will actually be required to meet these requirements is to go to the measure specifications and deconstruct the quality measures into their unique data elements; and that's what we did.

HCI: Was that difficult to do?

Metzger: Yes; and I don't know anybody else who's done it. And to do so, we had to go to each quality measure and its specifications, as well as the underlying data dictionary. Then we had to build a software program that allowed us to manage the information, to get the unique data elements associated with each measure. Because, for example, the stroke measures share a lot of data elements, so that the diagnosis for ischemic stroke, for example, shows up in many measures. So we had to figure out, what is the unique set of data elements involved for this set of measures? Every hospital has to get every unique data element from their EHR -that's what meaningful use is. It also says they have to confirm that the calculation and reporting were performed by certified EHR technology. The inpatient EHR is a bunch of applications, even if from one vendor-the order management piece, the CPOE piece, the registration piece-and it varies from vendor to vendor as to how that portfolio will be carved up. So we went in and figured out the likely source of each data element. Then we said, let's look at the subsets of data, to determine which ones will be easier to do, and which will be harder to do.

So the first question is, what will the easier data be? And 99 percent of hospitals have registration/ADT [admissions-discharge-transfer] systems, because they have to, to do billing. And another really easy type of data is laboratory results, because everybody has that online. Now, it will still have to be captured in the EHR, but at least they've got the data. So if you have those two kinds of systems, you have 18 percent of what you need for the measures-not a large percentage. So you've got those systems. In addition, one big change to the final rule [from the interim rule] is that meaningful use does include the emergency department, as do the measures.

HCI: Just to give readers one example, consider what's involved in facilitating the reporting on one of the 15 requirements, which is “number of patients who have received VTE prophylaxis or have documentation of why no prophylaxis was given the day of or the day after the initial admission or transfer to the ICU or surgery end date, for surgeries that start the day of or the day after ICU admission or transfer.” Ostensibly, it should be relatively easy to find and report out this measure, but in fact, as you point out in your report, there is a very complex set of inclusion, outcome, and exclusion data elements, spread out across multiple clinical information systems, that have to be gathered successfully in order to report this “single” measure; and the care settings include the ICU and the OR, and possibly the ED.

Metzger: Yes. It's complicated by a number of factors, including the fact that there are several kinds of VTE prophylaxis. And there is considerable complexity around surgical processes and timing as well, in the data elements for this measure. What's more, there are two kinds of VTE prophylaxis treatment, pharmacologic or mechanical; and look what the measure introduces about the timing, and over questions such as whether the patient is on a clinical trial relative to VTE, or whether the patient was admitted for comfort measures only. Further, is the patient an obstetric patient? Does the patient already have VTE? Has the patient already had a stroke? Or, if the patient is admitted for certain types of surgery, certain procedural codes will be excluded. The fact is, people who don't spend day in and day out collecting these measures simply won't realize how complex this is. It's not just, “Did we do VTE prophylaxis?” And the person who administers this is a nurse; so the outcome would appear within nursing documentation, which, by the way, is not a part of meaningful use.

HCI: So the drill-down leads to complexity everywhere.

Metzger: Yes, and it's in the expression of the measure, and to which patients each measure applies.

HCI: So the bottom line is that the work involved in collecting every one of these 15 quality measures is much more complex than people have realized so far?

Metzger: Right. And then you think that VTE prophylaxis is one thing, and it's much more than that. And then some of these things involve physician documentation and some don't.

Now, going back to the logic of the paper, let's say I'm a hospital that has registration, ADT, and lab systems; that gets me 18 percent of the way. And then if I meet all the explicit data capture requirements for meaningful use Stage 1, how much better off am I? Well, the answer is that you've only captured 17 percent more of the data elements, so you're now up to a “whopping” total of 35 percent of what you need. So when you do CPOE for meds, allergy lists, and problem lists, and vital signs (though vital signs don't help you with those measures), you're up to 35 percent of what's required. Among other systems required to make this work are CPOE beyond medication orders, physician documentation, other nurse documentation, discharge instructions, medication lists, and other documentation (physical and occupational therapy).

HCI: Those systems, along with the eMAR, provide up to 64 percent of the data needed to move forward successfully in Stage 1; that is a “wow.”

Metzger: Yes, I think that is the “wow.” It says that people had better start thinking about these quality data elements right now.

HCI: Some CIOs are moving forward on basic CPOE and basic physician documentation, but many have not even begun to move forward yet on eMAR. But it's clear that your analysis shows they do need to move forward on eMAR and the more advanced physician documentation-notes, etc., right?

Metzger: For the physician documentation, I can imagine developing a very basic patient face-sheet, not really the full notes, which might bridge that gap. And sometimes, you can embed data capture into a condition-specific order set. I know there are CPOE systems that allow you to include in the order set questions with pull-down answers. So for the VTE prophylaxis measure, for example, the physician would have the opportunity to say why they're not ordering VTE prophylaxis, which is one of those troublesome data elements; so there are a few ways to get around that problem for the moment. But a lot of systems aren't set up to do that. But I don't know how you do the meds administration part without eMAR.

HCI: So in other words, if they don't have eMAR, they won't even be able to capture the measure data elements they need?

Metzger: That's right, because meds administration has to be done and documented by nurses; the specifications are clear. So the way I read this is, it isn't some quality nurse coming in after the fact doing the electronic documentation. So that's where the combination of the specifications and the data dictionaries sheds important light on these measures; which is why we took the approach we did. Another issue revolves around your systems in surgical suites; we found that nine percent of the data elements involve care actually delivered in the surgical suites. And even when we did this analysis for the interim rule, 14 percent of the data came from ED systems, and that's true in the final rule also. And in some hospitals, 75 percent of patients are admitted through the ED. The challenge, of course, is that many hospitals that have true clinical OR systems, not just scheduling systems, and have ED systems, generally have best-of-breed systems in those areas that aren't from their core clinical vendors.

HCI: So the huge implication in all this is that if you don't have CPOE beyond meds orders, and advanced physician documentation, if you don't have eMAR, if you don't have ED and OR and have those systems linked to the core EHR, you won't get there?

Metzger: That's right. And everybody's been focused on that table [laying out the general requirements in the final rule]; and they haven't drilled down on the one line item for quality measures.

Healthcare Informatics 2010 October;27(10):40-42


The Health IT Summits gather 250+ healthcare leaders in cities across the U.S. to present important new insights, collaborate on ideas, and to have a little fun - Find a Summit Near You!


/article/complexity-behind-quality-measures

See more on

betebet sohbet hattı betebet bahis siteleringsbahis