Electronic health record systems were not designed for population health, but help is on the way.
Just when physicians thought they had electronic health records (EHRs) under control, they must now learn how to use other health IT applications that support the shift to a new kind of healthcare delivery known as population health management (PHM). Physicians nationwide don’t really have a choice if they want to get paid what they’re worth in the future.
Over the next few years, value-based reimbursement is expected to replace a large portion of fee-for-service. Physician practices that aim to succeed under this new payment model-especially those that plan to take on financial risk-must learn how to manage population health.
In other words, they must keep their patient populations as healthy as possible so as to produce better outcomes at a lower cost-the definition of value. While human interventions are needed to achieve this goal, information technology also has an essential role to play in PHM.
Medicine has always been about information, but the amount of data needed to manage the health of a population is a quantum leap above the volume of information contained in the typical patient record. To start with, physicians need data not only on what has been done for each patient within their practice, but also on the services provided to that patient in other care settings. They must also be able to monitor the health and compliance of patients between visits. They need to understand the health risks of particular subpopulations and the financial implications of those risks.
Providers must also be alerted to the care gaps of individuals and must have a well-organized system to close those gaps. Moreover, organizations must measure their performance on quality, efficiency and patient experience measures, as well as the performance of individual sites and providers.
Smaller practices are expected to merge or form joint ventures to meet the challenge of value-based care and acquiring the IT infrastructure needed to support it. Some observers expect that most small- and medium-sized practices will eventually join larger organizations, such as independent practice associations (IPAs), accountable care organizations (ACOs), and bigger medical groups
The Medicare Access & CHIP Reauthorization Act (MACRA) will accelerate this movement when it takes effect in 2019, predicts Lawrence Casalino, MD, professor of public health and chief of the division of health policy and economics at Weill-Cornell College in New York. “There are going to be howls of outrage, but MACRA is going to move physicians much faster into organizations that are capable of doing value-based care,” he says.
Most of the applications required for PHM are too expensive for small- or medium-sized practices. In fact, only organizations that include at least 500 to 1,000 physicians can afford sophisticated analytic software, says John Moore, founder and CEO of Boston-based Chilmark Research, a health IT research firm.
EHRs were not designed for PHM, but the leading EHR vendors have begun adding PHM modules to their offerings and are developing more (see EHR Solutions sidebar on page XX). Thus many physicians are looking to their EHR vendors to supply what they need.
Moore is skeptical about the value of EHR-based PHM applications. But he admits that these solutions can be helpful to some extent, “especially in the ambulatory physician market.” For example, he says, EHRs can help physicians with quality measurement and reporting.
A wide array of health IT solutions have been devised for PHM. Besides the EHR-based PHM modules, there is a plethora of standalone solutions, each of which addresses a particular part of the PHM puzzle. Sorting out this jumble of applications can be bewildering, and getting them to work together may require health IT experts who are available only in large organizations.
To help understand the core IT functions needed for PHM, here are a few basics.
A patient registry provides a central database for quality improvement and PHM. A registry shows when a person was last seen by their healthcare provider, what was done for them, their health status at the last contact, their latest lab results and when the patient is due to visit again. When combined with best-practice protocols, the software can be used to alert providers and care managers to care gaps. A registry can also be combined with automated outreach applications to alert patients when it’s time to make an office appointment. And a registry can be used for quality reporting and performance measurement.
Analytic software can be applied to registries or, in large organizations, to data warehouses for a variety of purposes. For example, many healthcare organizations use risk stratification tools that segment the population by health risk-i.e., the risk of a patient becoming sick or sicker.
Risk stratification, which is used in care coordination and financial risk management, enables organizations to plan how best to take care of their patient population. They can assign care managers to look after high-risk patients, make sure that people with mild or moderate chronic diseases receive recommended care and educational materials, and make robocalls or send letters or online or mobile alerts to healthy people who are due for preventive care.
The effectiveness of PHM activities depends on the reliability and comprehensiveness of the data that these applications use. The three main types of data used in PHM today are clinical, administrative and claims data. (Patient-generated data are expected to become important in the future.)
Clinical data are rich and timely but are limited because of the lack of interoperability among health IT systems. Also, the majority of clinical data are not structured; they are trapped in free text and unavailable to analytics. In contrast, billing and scheduling data from practice management systems are structured. So those data are easier to use than clinical data, but are not as detailed or actionable.
The most comprehensive type of data is claims data, which covers all of the billable services provided to patients across all care settings. But claims data are fraught with errors and are usually a few months old when providers see the information, making it much less actionable than clinical data. Moreover, health insurers do not provide claims data to most healthcare organizations, notes David Nash, MD, dean of the Jefferson College of Population Health at Thomas Jefferson University in Philadelphia.
ACOs that participate in the Medicare Shared Savings Program (MSSP) can obtain Medicare claims data, Nash adds. But interpreting that data can be very challenging unless a group has a health IT expert on staff, he says.
Some ACOs use commercial analytic tools to crunch Medicare data, as well as claims data from private payers, if they can get it. Some EHR vendors also provide analytic software. For example, Emerald Physicians, a 50-provider group in Hyannis, Massachusetts, uses a module attached to its eClinicalWorks EHR to analyze Medicare data. But even with the help of this tool, “the amount of data coming in is immense,” says nurse practitioner Susan Harrington, the group’s clinical IT director. “It’s a process to sort it and decide how you want to view it and manage it.”
While Emerald Physicians has figured out how to analyze population data, many other practices will find it beyond their capabilities. Moreover, some groups may view the process of hiring and directing care managers as too complex and expensive. So some EHR and PHM software vendors offer outsourcing options.
Observers are skeptical, however, about the viability of outsourcing care coordination. Moore points out that outsourced care managers are not based in the community. “They’re going to be in some town halfway across the country, and it’s going to be whoever happens to have that person on their roll call that day. There won’t necessarily be the personal kind of relationship that a person has with their physician or that practice or has met the care manager in that practice,” he says.
As mentioned earlier, Moore has doubts about the viability of EHR-based PHM solutions, mainly because they’re physician-centric and are limited to particular practices or other care settings. Most EHR vendors, he says, are unable to aggregate data from disparate EHRs and share the information across care settings.
Cerner is one vendor that can do it, but is selling its HealtheIntent platform mainly to large organizations. Allscripts also has capabilities in this area, having acquired a health information exchange (HIE) vendor called dbMotion. But eClinicalWorks cannot exchange information outside its EHR, he says.
Additionally, Moore says, “We’ve found the registries provided by the EHR vendors are woefully inadequate, by and large. Cerner has done a really good job, and Epic has a good registry. But the vast majority of ambulatory EHR vendors have not done a good job with registries.”
David Nash of Thomas Jefferson University agrees with Moore. “There’s way more heat than light being generated on the health information side,” he says. “All of the big players-Epic, Cerner and so on-are hurriedly putting together their population health solutions and strategies.” So far, he adds, there is no peer-reviewed evidence as to how well these systems work.
The practice representatives we interviewed have not seen the interoperability the vendors say they’re building into their PHM solutions. The Allscripts Enterprise EHR, for instance, doesn’t give New West Physicians, a 65-doctor group in Denver, Colorado, the ability to obtain community data, says Ken Cohen, MD, the group’s medical director. Susan Harrington of Emerald Physicians says that eClinicalWorks is now connecting the group to its private health information exchange, but that the HIE will only enable them to trade data with other eClinicalWorks customers.
Some physicians are also less than impressed by EHR vendors’ registries and have found them lacking in comparison with standalone applications.
Gretchen Hoyle, MD, of Twin City Pediatrics, a seven-doctor practice in Winston-Salem, North Carolina, led the effort to build a registry for her group several years ago. The doctors in her practice agreed on a set of guidelines and decided which fields to include in the registry, which they used to track patients with chronic conditions. They didn’t interface the registry with their EHR, so their staff had to enter patient data twice.
Despite that drawback, Hoyle says, the homegrown registry is superior in some ways to the one in the Epic EHR that the practice has used since it was acquired by Novant Health a few years ago. “There are some things we can’t do with Epic that we used to do and still do [with our own registry],” Hoyle notes. “We can run a list of everyone who’s due for a checkup, a list of everyone who has a certain diagnosis, or everyone who has two diagnoses and is on certain medications.”
One problem with Epic, she says, is that the physicians can’t create their own registry fields. Also, she notes, the Epic registry is based largely on billing codes, so it can’t distinguish, for example, between a child asthma control test and a visit for developmental screening and checkup. “Epic is way better because you don’t have to do dual entry of data, and you can search more people and won’t miss anybody. But there’s a lack of flexibility,” she says.
Similarly, family physician Jennifer Brull, MD, of Plainville, Kansas, found that her
e-MDs EHR lacked a robust registry, which she and her four colleagues wanted to use in their quality improvement activities. The practice purchased a standalone registry with funds from a health plan that was promoting patient-centered medical homes.
The outside registry worked pretty well after it was connected to the EHR; but later the registry’s vendor reprogrammed it to use claims data rather than clinical data. That was not helpful, Brull says, so the physicians stopped using it.
“We miss the old registry,” she remarks. “To bridge the gap, we’re running reports in our EHR, and it’s a much more time-intensive process. We can get the same data, but it takes a lot more time.”
Some physicians have had better experiences with their EHRs. New West Physicians’ Cohen says that while the Allscripts registry is far from ideal, it is adequate for the needs of the group’s physicians. “I can sit at my desk and open up my entire population of patients with asthma, hypertension or diabetes,” he says. “I can look at measures of control and measures of quality.”
The group distributes quarterly reports to its physicians and midlevel providers on the care gaps of individual patients. The reports also pinpoint patients who need help, such as people on warfarin who are outside the therapeutic range, he adds.
Similarly, Christopher Berard, DO, an internist in a three-doctor practice in Babylon, New York, says his athenahealth EHR enables him to run quarterly reports to improve the quality of care. There is a tab in each patient’s chart that alerts him and his colleagues about the patient’s care gaps, and those alerts can be customized.
To find out what is happening with his patients outside of his practice, Berard can look at Surescripts for medication histories, and he receives some information online from his hospital. But only the ACO he belongs to, Long Island’s Beacon Health Partners, has comprehensive data on the care that his patients receive-and that is available from only some payers. The ACO uses the Crimson analytic program from the Advisory Board Company, a healthcare consulting firm. The software crunches the claims data that the ACO can access.
According to Simon Prince, MD, the ACO’s former medical director, the Crimson application is used partly to provide feedback to the doctors in the ACO about their clinical performance. In addition, the ACO uses the software to identify the highest-risk, highest-cost patients so that its centralized care management team can intervene with them.
Because the claims data is not current, the ACO asks the doctors to “scrub” it, says Prince, to weed out patients who have died or are no longer in their practices. The physicians can also refer patients whom they deem high risk to the care managers.
Health IT for population health management presents physician practices with a conundrum. Unless they’re in large organizations, they can’t afford most standalone solutions. On the other hand, the PHM software offered by EHR vendors has some serious drawbacks.
Where does this leave the small- or medium-sized practice? Physicians in such practices should take a careful look at the PHM applications offered by their EHR vendor, as well as the functionality already in their EHR that can be adapted to PHM. They can use that as a starting point, and gradually expand their capabilities from there.
“You can’t go from zero to 60 immediately,” notes family physician Brull. “You have to start with things that are easy and understandable and available.”
Hoyle, the pediatrician, points out that physicians should not be afraid to make less- than-ideal choices. “You never really lose by trying to move forward. Even if you pick the wrong solution or it’s not your ultimate solution, there are so many things everybody learns by doing it.”