MedStar Health, a regional healthcare system with a network of nine hospitals and 20 other health-related businesses across Maryland and Washington, D.C., has been using GE Healthcare's Medical Quality Improvement Consortium (MQIC) to report and analyze quality measures. Peter Basch, M.D., MedStar Health’s medical director of ambulatory EHR and health IT policy, spoke to HCI Associate Editor Jennifer Prestigiacomo about the exciting new project his system is undertaking to identify quality measure outliers to globalize best practices.
Healthcare Informatics: When did you start using GE Healthcare's Medical Quality Improvement Consortium (MQIC)?
Peter Basch, M.D.: We’ve been using MQIC for a good number of years, and we have begun to be more robust MQIC users in the last year or two. Our health system has matured use of the EMR [electronic medical record]; and our doctors and staff have been much better in entering information and structured data, and thus engendering in our end users a deeper sense of trust and reliability in the reports. One of the common experiences of clinicians in getting reports on [clinical] data is to discount it in the sense of, ‘well this doesn’t even look remotely correct.’ We’ve had a different experience with our MQIC reports, as it is populated with our clinical data. And as mentioned before, our doctors are better about reporting things that need to be reported as structured information. The reports that you can deliver via MQIC now are much more reliable, and thus trusted and looked at as useful to our clinicians and our managers. We look at our MQIC reports as our regular, sometimes annual or quarterly reporting, in terms of helping to drive our quality improvement agenda.
HCI: Are you using MQIC in all nine hospitals and all physician practices?
Basch: The use of MQIC right now is really dependent on which clinicians we have using the EMR and what particular reports are relevant to those clinicians. So, at this point our primary users of MQIC reports are particularly primary care and a couple of medical subspecialties like endocrinology and cardiology. That does not cross over into all of our hospitals. Five hospitals and one non-hospital entity— our primary care group—have clinicians that use the MQIC reports.
Prior to us going electronic we had a group within MedStar, which still exists, called the Ambulatory Best Practice Group, comprised of representatives of each of our MedStar units that conduct outpatient care. And the work of that group up until recently was essentially commissioning and conducting clinical reporting. And it would be done by manual chart abstraction. So the entire year of work for that group would be to look at all the metrics we want to look at, take 40 or 50 charts per clinician, and spend a year to do the chart abstraction. At the end of the year, we’d look at our reports, and compare it to where we were the previous year.
Since going electronic and our robust use of MQIC, the decision of our reporting is typically done in our first meeting, and then the running of the reports, broken down by practice, and by individual provider, takes maybe a couple of hours. So what we used to do in the course of a year is now done almost in real time. We spend the rest of the year looking at what we can do to improve care, or which operational measures can be applied and how they are working in practices instead of in theory. So, it’s really changed the focus of the group.
HCI: Are you currently tracking all 15 quality measures specified by the meaningful use Stage 1?
Basch: I assume it’s going to primarily be an offering through MQIC that’s available to us, and we have other reporting tools as well. But we look at the MQIC reporting as giving us clear guidance at a rolled-up level, so we know as a health system, a hospital entity, and a primary care practice group where we are in aggregate. Do we have e-prescribing implemented in such a way that our docs just don’t get it?
As we know meaningful use is done provider by provider. The other part of what MQIC will give us is individual provider reporting so that doctors will be able to look at all their measures for which we made a system determination on where they stand. I think one of the strengths going forward in 2011 with the product is providing that information to every provider using the EMR. We think that that near real-time feedback will show providers where they are on measures. I think that it’s particularly unhelpful with other vendors that say they’ll submit your measures for you. Good luck knowing where you stand. As we know meaningful use is all or none, so you really do want to know early on where you are with performance. And if you’re not doing well, show me the information that I need so I can see where I’m not doing well, and perhaps with some tips as to how I might change what I do in practice style, or prescribing style, or appropriate inputting of structured information, in such that I’m able to improve.
GE has made a commitment in their version 10, which we’ll be getting in first quarter of 2011, to include metrics performance for meaningful use measures. Right now we’re looking at other quality reports through MQIC and PQRI [Physician Quality Reporting Initiative] reports.
HCI: In the CSC report about the complexity around quality measures, it noted that hospitals really have to have more than just medication orders, to achieve meaningful use. Where does MedStar stand on that?
Basch: I think getting all those measures addressed in a timely measure is going to be challenging for most hospitals, including ours. One of the things I feel good about is in the outpatient environment, I think most of our docs are used to putting in their problems, meds, allergies all as structured data, ordering medication properly, e-prescribing, providing patient information in a timely fashion. We look at meaningful use in the outpatient arena as a rung in a ladder toward a journey that is more than meaningful use.
For example, meaningful use says, put in one problem list as structured data. Our internal minimum use requirement says, put all your problems in as structured data, and update it regularly. So we’re already asking way more of our docs than what meaningful use does. In that way we see meaningful use as a standard step along the way—and that’s not to minimize the challenge. It will be far more challenging to meet inpatient meaningful use than meaningful use for eligible providers. We’ve had a full clinical system for our clinicians for a couple of years now, whereas in the inpatient we haven’t implemented CPOE yet.
HCI: Can you tell me which quality measures you are reporting and why you chose these?
Basch: The measures we’re primarily focusing on right now include cancer screening measures and diabetes, asthma, heart failure, and coronary artery disease—the usual suspects in terms of what most people look at—where we spend a lot of money on healthcare, and where we get less than stellar results. We’re all pretty much focused on the same ones.
HCI: How has tracking these quality measures changed the quality of care at MedStar?
Basch: One of the hard realities when you look into the mirror is we do a good job much of the time, but things do slip through the cracks, either because of multi-tasking, fatigue, inattention, patients not showing up for requested follow-up, and so forth. So when we look in aggregate, what we do with our patients, our performance is often lower than we would think it to be, and certainly not always as good as our patients deserve. So when you begin to look at reporting you can trust and that’s usable, you can begin looking at the brutal truth of what actually exists. You can inch that forward and say, ‘we’ll here’s what I can do to do better.’
Now we’ve looked at metrics at preventative and diabetic care over the past two years and found approximately 5 to 10 percent improvement since we’ve begun to pay attention to the reports. There are a couple of other things that reports allow us to do, and one of the more interesting projects we just started is if you look at your reports, and not just look at who’s doing a little better, but you can now look at who are your top performers. Let’s say for example, the NCQA [National Committee for Quality Assurance] or baseline for a particular metric is achieving 40 percent of your patient population having a particular measure, and certainly you want to make sure all your providers make it past that baseline. We might have a couple of providers who perform two to three times better. To me, an interesting use of those reports is to identify who those people are, and then through either observation, or survey, or conversation find out what is it that these providers are doing differently about their practices, or how their practice interoperates with patients, or how their patients perceive their own health, and so on. So, that’s a really interesting way to looking at reports just for quality improvement, but for quantum quality improvement.
HCI: So you just started this new project?
Basch: We just unveiled this at a recent MedStar patient safety conference and presented this idea of this as a unique way of looking at quality measures. Instead of looking for the poor performers and helping them improve, let’s look at our top performers, and glean what they’re doing and be able to generalize it across the health system.
HCI: What has the most eye-opening finding been since tracking data for quality measures?
Basch: I can give you one story of a learning that came as a surprise to me. We’ll go back to the example of diabetes. So one of the commonly accepted measures in diabetes is to try to get the patients who are most poorly controlled—we’re using a hemoglobin A1C measure of greater than 9.0 to be less than 15 percent of one’s diabetic population. So, in other words, we know we can’t get everyone into good control, but on average, get at least 85 percent of them with a hemoglobin A1C of less than 9.0. In the health system we’re real proud that we just fit into the threshold of that metric; I think MedStar Health as a whole is at 14.3 percent of our diabetics are poorly controlled. And that’s good.
So, what I did using reports is looked across a subset of our population where we had very granular provider reports on diabetic populations, and of these docs, one provider in particular had only 4 percent of her diabetics in poor control. When I asked her how she did it, she couldn’t answer. When we were looking across the diabetics in poor control, we were looking for non compliance, level of obesity, insurance status—all the usual things we thought it would be, didn’t hold. One of the commonalities of all those patients in poor control was of complete surprise; none of them were on insulin. There were all on oral agents, diet, and exercise. With all the great orals out there, a lot of clinicians feel uncomfortable with insulin now. Patients don’t like to give themselves injections, so they’ll say ‘I’ll do better next time.’ So we end up with a poorly controlled diabetic for months and years, and never have their sugar under control because the doctor is not putting their foot down early enough starting insulin.
When I talked to this doctor who had the great stats, she said ‘oh yeah, I should have told you. When I have a diabetic that’s way out of control, I give them two months, and if in two months they haven’t made significant improvement, we start insulin. And I tell them, if you can lose that 100 pounds over the next year or whatever it is, or demonstrate to me six months from now with good sugars that you’re actually following a good diet, we can always stop the insulin.’ What she’s doing is actually being more aggressive in getting the sugars controlled quickly, and once the patient shows they can manage their sugar well, and possibly taper off insulin—the reverse of what a lot of people do by exhausting all other opportunities first and going to insulin as a last resort. What that means is that we have lots of patients with out of control blood sugars for long periods of time. So that’s one learning of the data.
We’re looking to see if we can get similar learnings. If we stick with diabetes, how do you get more than 50 percent of your diabetics vaccinated for pneumococcal disease or how do you increase compliance with eye exams?
HCI: What other performance findings have you made across the health system?
Basch: Another thing we found, with the ability to look across entities, as an aggregate, we do pretty well. But when we break it down to the individual, where changes can actually be made, a startling thing was that in the same business unit, you have providers ranging from dismal to stellar. I know the doctors don’t range from dismal to stellar. We have the ability to look at the data and ask if they’re dysfunctional EMR users? Are they documenting somewhere in their note that something is happening, but not somewhere that can be seen in objective reports? Or is it something about their practice style? Are they not addressing the metrics that we as a health system embrace? That was a surprise because I was expecting to see a lot of uniformity in metrics.
HCI: What are some of the things you’re looking forward to doing in the future with quality reporting?
Basch: The thing that we’re looking forward to now is getting that kind of clarity and acceptance of the report view of oneself. Clinicians have to take a good long look in the quality mirror, and if there’s a problem with the data, then let us know, so we can see if we’re not capturing something appropriately. The willingness to use it as a guide will move toward better quality. If there’s not that trust, then it’s really hard to use information to guide you in future care.
Another thing we’re doing now is starting with the basics. Certainly, there are a lot of sophisticated things that can be done, but I think if we should start with the basics: cancer screening and chronic care management for diabetes, heart disease, asthma. We can then begin to pull all of our providers, who are not performing as well, higher, and do that relatively quickly.
A third thing I look forward to over the next year is scanning our data for who are our outliers and top performers. And work with them in their practices and with the data to figure out if we can find other examples, aside from the one I talked about with diabetes, of changing practice style, engaging patients differently, or using the practice unit differently. [This will] help us not just push up from the bottom, but to pull from the top, to see if we can globalize a best practice, so that we can in a relatively short time see within our health system a sea of top performers.