Data Analytics Offer Tantalizing Opportunities for Practice Improvement
Practice Management Feature — April 2016
Tex Med. 2016;112(4):37-41.
By Joey Berlin
Physicians aren't shy about voicing their opinion of electronic health records (EHRs): They're frustrating, time-consuming, and detrimental to patient care. And federal meaningful use requirements are incredibly burdensome, with "meaningless abuse" an increasingly popular alternative term.
EHRs and meaningful use, for the time being, aren't going anywhere. And as more beleaguered physicians grudgingly enter patient data, they're creating a well of statistics that can do so much more than sit in a program and consume digital space.
Properly leveraged patient data are valuable, and physicians and groups are already making effective use of them. As medicine moves away from fee-for-service payment models and toward value-based payment, "analytics is really at the heart of being able to do that management," said Houston infectious disease specialist Tobias Samo, MD.
Data from the past can help improve the present. Predictive analytics are the future.
"We can tell who's most likely to be hospitalized within the next three to six months," says Stephanie Kreml, MD, chief medical officer of the Austin data analytics group Accordion. "The associated cost predictions for that and the specific diagnosis group related to that event can help care coordinators be more proactive in reaching out to those patients so they can hopefully prevent those events from occurring."
Dr. Samo says he's been involved with health information technology (HIT) for more than two decades. He served as medical director of information technology at Houston Methodist Hospital for four years, then was chief medical officer at Allscripts for six years. He now does independent HIT consulting in addition to his part-time clinical practice.
The shift to fee-for-value payment models requires technology that can turn data into information, Dr. Samo says. While straight reporting of patient data represents part of that equation, fee-for-value will also require predictive analytics, such as performing a risk stratification of patients to identify which ones are at the highest risk for requiring additional support and resources.
"All of that is impossible to do without having [EHR] data in addition to claims, lab data, probably socioeconomic data, as well," Dr. Samo said. "There’s a lot of work going on with analytics that goes well beyond meaningful use."
Plano family physician Christopher Crow, MD, is the president of Catalyst Health Network, an accountable care organization (ACO) of independent physicians. He agrees supplementation of a practice's EHR data is essential.
"It's important to marry what's happening in your practice with the claims data because the claims data have most of what's happening to that patient across the health care delivery system in that particular area," he said. "[For] some of that stuff, your [EHR] is going to be incomplete."
Examples abound for how data from medical visits can yield valuable knowledge. Using EHR data, and potentially claims or other outside data that might be available, Dr. Samo says a practice can identify patients who are "frequent flyers" to the emergency room.
"If I identify 10 patients who have been in the emergency room five times or more, then I can now get my case managers to get in touch with these patients and find out why are they coming in, why are they using these services, and what's the way that we can reduce that frequency," he said.
He adds "huge opportunities" exist to "identify those types of outliers, and then through whatever it may be — case management, maybe it's education of the physicians, education of family members, etc. — you go from that population view down to the individual patient view and figure out ways to improve their care, so they're using the appropriate venues of care."
Jennifer McClung, director of systems integration for Stratifi Health, which provides information technology services for Catalyst and other independent physicians, says many physicians she's worked with will make an assumption about a trend in their practice based on what appears to be a gut feeling.
"Sometimes it's [because of] whatever patient they saw in the last exam room," she said. "When you actually step away from the emotion and go back to data and facts, we actually have [a different] situation that needs to be looked at."
Ms. McClung says she sees a shift toward predictive analytics. Accordion is one company participating in that shift.
Accordion has developed predictive models for insurance plans and ACOs to identify patients who are most likely to experience acute events or need hospitalization. The company eventually wants to extend its products to smaller practices. Dr. Kreml says Accordion is mainly using claims data but also ties in lab results and is beginning "to tap into EHR data to provide a fuller picture of what's happening with the patient population."
Accordion's predictive modeling process involves pulling out the most relevant pieces of data. The process uses algorithms and regression analysis to develop a model that can predict a certain type of event for a patient, such as hospitalizations.
Once that model exists, Dr. Kreml says, there are a couple of different ways to set it up. One is to allow the physician to enter specific parameter values for a given patient, similar to the Framingham Risk Score that calculates a person's risk for cardiovascular disease over the next 10 years.
"We can put that into the model to then have an output of specifics about cost and outcomes," Dr. Kreml said. "And in fact, we have developed something really similar to that where it's a form and you can [enter values] specifically for total knee replacements ... or, we can take all of the really rich information in the claims data, everything that we used to develop the model. We can take those data feeds and put that into the model to then come up with a specific [prediction] within the next six months: This particular patient has this probability or this likelihood of having this event, and here's the expected costs associated with that, with this particular confidence level or probability."
Dr. Crow says Catalyst has around 90 locations and 300 primary care practitioners in its network. Catalyst gets its data from insurance carriers, analyzes it, and reports its findings at the practitioner level. It examines cost and utilization reports for indicators like emergency room usage, hospital usage, and generic prescriptions.
"It's pretty simple — just the way you had report cards growing up," Dr. Crow said. "You show people where there's opportunity, and they go from there."
At Village Health Partners, a Catalyst network group where Dr. Crow has been practicing since 2004, "You can see much improvement around any of those quality metrics, whether it's mammograms or colonoscopies or diabetic measures, hypertension, generic [medication] usage, because they're now aware of it and they're thinking about it more."
The analytics of medicine are subject to change. About a decade ago, when Dr. Crow was in a three-physician family medicine practice, it made adjustments based on what he says was a "metric du jour" at the time: It used data from its EHR system to determine what percentage of its diabetic patients were getting regular eye examinations to scan for diabetic retinopathy.
The physicians found out just 36 percent of patients were getting the regular eye exams. As a result of those analytics, Village Health purchased the eye equipment necessary to perform the exams and increased its eye exam rate to 60 percent and implemented a long-term program to raise the exam rate to 90 percent.
Today, Dr. Crow says that particular metric isn't one he uses much anymore because eye problems for diabetic patients are a lagging indicator. "If your eyes are having diabetic problems," he said, "then you're about 10 years behind."
"But the principle of it is the same: If you just start measuring someone for the first time and you've never done it before, most physicians would think they're above average in these categories," Dr. Crow said. "But by definition, everybody can't be above average, so once you start doing the data analytics, you start to be able to tease that out."
More than 90 percent of health plans in the United States use the Healthcare Effectiveness Data and Information Set (HEDIS) for quality measures.
EHR Usability Still a Problem
The wealth of data EHRs house and the potentially spectacular applications of that data do nothing to obscure a problem physicians everywhere continue to curse: Some EHRs are about as usable as a submerged iPad.
A recent University of Texas research project found many EHR vendors didn't know how to design usable software, and vendors will require input from other stakeholders, including physicians, to improve the usability of their products. (See "Better Bridges, Better Systems," September 2015 Texas Medicine, pages 39-43.)
"This is in some ways one of my pet peeves: I don't see anybody being really innovative in the usability area," Dr. Samo said. "I think there are lots of opportunities there for innovation when it comes to usability."
Dr. Kreml hears her share of "complaining and griping about EHRs," and knows in following CMS' requirements for meaningful use, a physician is "tracking things that may not necessarily be relevant to the outcomes that you care about."
"I totally understand the frustration, and I don't know if our being able to provide a toolkit for a doctor out there in a solo practice or in a small group … if that's enough of a motivation to say it is worth it to deal with all the headaches on the front end and capturing all the data," Dr. Kreml said. "But maybe it's one piece that could help."
Most small practices that aren't in an ACO don't have the same access to outside data, or the personnel and infrastructure, to make use of it, Dr. Crow says.
"If you're part of a large system, like Seton [Healthcare] in Austin, then that's going to be provided for you," Dr. Crow said. "But if you're part of a smaller practice, you're not going to probably have much of that. And if you did have it, you don't have context. You don't have the ability to compare yourself as easily across the market. The great thing about our ACO is we compare across the market to each other and across the nation into other ACOs."
Most EHRs don't pull in claims data, Dr. Samo says, and that's one of the factors that will encourage small practices to move into larger groups, both physician-owned and hospital-owned. He says "that larger infrastructure is going to be something that's difficult for the smaller groups and solo practitioners to be able to manage. They'll be able to do some of it with their [EHR], but certainly, once you start getting into risk-sharing, etc., then that's really going to be beyond the capabilities of those small groups."
TMA PracticeEdge helps bring physicians together to form ACOs and gives them the leverage and resources to participate in value-based care. TMA PracticeEdge helped Texas physicians form five new ACOs during 2015.
Although physician complaints and reservations about EHRs and meaningful use are likely to continue indefinitely, Ms. McClung senses a change in attitude that's paving the way for an analytics revolution.
"A lot of these things are possible because I see more and more of the physicians I work with increasingly open and receptive to that data," she said. "They want that information. They're embracing it, rather than viewing it as somebody giving a report card on them or judging them. The providers who are embracing it and going, 'This can really help me be a better provider,' are doing amazing things with the information."
Dr. Samo concedes some EHR features and problems "slow you down" but says the programs also can be excellent data collectors. And whatever the future of the CMS meaningful use program may be, he says the program has been an important part of encouraging practices to adopt health information technology. (See "Meaningful Use Program Still in Effect.")
In a less scientific form of predictive analysis, Dr. Kreml says she sees days ahead where practitioners are more receptive to generating and using mountains of data.
"We're going through a transition period right now, and I'm optimistic that eventually we'll get to the point where physicians on the front lines will see the benefit of being able to capture everything that can be used in a way that will make their lives easier and make it so they can sleep at night and not worry about, 'Does this patient fall through the cracks?' But it's going to take a lot of work to get there."
Joey Berlin can be reached by phone at (800) 880-1300, ext. 1393, or (512) 370-1393; by fax at (512) 370-1629; or by email.
Meaningful Use Program Still in Effect
Centers for Medicare & Medicaid Services (CMS) Acting Director Andy Slavitt sounded alarm bells in January when he told an audience at the J.P. Morgan Annual Health Care Conference, "The meaningful use program as it has existed will now be effectively over and replaced with something better."
Mr. Slavitt said more details would come in the next few months but mentioned general tenets of CMS' planned approach, including rewarding practitioners for good patient outcomes, allowing practitioners to customize their goals, allowing technology companies to build around the needs of individual practices, and developing initiatives for better health information technology interoperability.
The statement that meaningful use was "effectively over" generated everything from confusion to elation in the medical community. But Mr. Slavitt and Karen DeSalvo, MD, acting assistant health secretary for the Office of the National Coordinator for Health Information Technology (ONC), released a joint statement the following week clarifying the remarks.
Mr. Slavitt and Dr. DeSalvo noted the passage of the Medicare Access and CHIP Reauthorization Act [MACRA], which allows for consideration of quality, cost, and clinical practice improvements in determining Medicare physician payments. While saying MACRA "provides a significant opportunity to transition the EHR Incentive Payment Program for physicians towards the reality of where we want to go next," they noted the law still requires ONC to measure the meaningful use of health information technology under the existing standards.
"While MACRA provides an opportunity to adjust payment incentives associated with EHR incentives in concert with the principles we outlined here, it does not eliminate it, nor will it instantly eliminate all the tensions of the current system," they wrote. "But we will continue to listen and learn and make improvements based on what happens on the front line."
Back to article
HIT Survival Kit
TMA has created a Health Information Technology (HIT) Survival Kit to help members make the most of TMA's HIT resources.
The kit's seven modules guide users through understanding the current HIT environment and ongoing HIT compliance. The tool highlights TMA resources that can help physicians navigate every step. Each module contains items members can access free of charge as a benefit of TMA membership, as well as items members can purchase at exclusive rates.
April 2016 Texas Medicine Contents
Texas Medicine Main Page