In Focus: Implementing Health Information Technology--Lessons from the Medicaid Transformation Grants

Summary: States that received grants from the Centers for Medicare and Medicaid Services to advance the use of electronic health records and e-prescribing programs confronted challenges other states are likely to face as they implement health information technology initiatives using funds from the Stimulus Bill.

By Sarah Klein

Health information technology has long had the potential to transform the quality of health care by enhancing the ability to measure and analyze process and clinical outcomes—by condition, institution, and region.¹ But its widespread use has been hampered by a lack of funding, among other barriers. Congress has helped to address this through the American Recovery and Reinvestment Act of 2009 (ARRA), which appropriates $46.8 billion to encourage hospitals and providers to become meaningful users of electronic health records.² The funds are included in the ARRA's Health Information Technology for Economic and Clinical Health Act, or HITECH Act.

The incentive payments, which are designated for Medicare and Medicaid providers whose practices meet certain criteria for patient volume, will begin reaching those providers in less than two years—a short timeframe considering the amount of regulation and guidance that must be developed first. The federal government must determine what constitutes the meaningful use of such records, as well as standards for interoperability and security, and remedies for the misuse of information. That work, which is spearheaded by the Office of the National Coordinator for Health Information Technology, has focused on ensuring that investments in health information technology lead to measurable changes in the prevention and management of chronic diseases; a reduction in medical errors and racial disparities in health care; improvement in coordination of care and communication between providers and families; and a contribution to public health efforts.

States, which are charged with disbursing the Medicaid funds, also have a considerable amount of work to do to leverage those funds to have an impact on the quality of care. Perhaps the best evidence for this is another recent federal program that supported similar health information technology projects at the state level. The Medicaid Transformation Grants were issued in two phases in 2007 and provided $150 million to 35 states, the District of Columbia, and Puerto Rico to implement electronic health record systems and tools to enhance clinical decision making. (The grants also financed projects aimed at reducing spending on prescription drugs and preventing fraud, waste, and abuse in the Medicaid program.) The state initiatives ranged from the installation of e-prescribing systems for physicians in rural communities in Tennessee to the creation of a Web-based medical record for children receiving health care services through the foster care program in Texas.

To date, the states have accomplished a great deal, considering the relatively modest amount of funding they received. Arizona, which received $17 million, built a Web-based health information exchange known as AMIE that gives hospitals and providers access to hospital discharge summaries, medication lists, and laboratory information at the point of care. Developed as a proof of concept, AMIE contains information about Medicaid and non-Medicaid patients and will be expanded to include clinical data and clinical decision support tools. Minnesota, which received $2.8 million, expanded an existing Web portal to enable providers to see three years' worth of claims data on emergency department visits and inpatient stays for a patient, along with one year of medication history (the latter is combined with records of multiple insurers). The system was designed to help primary care providers identify individuals who were repeatedly hospitalized or sought emergency care and needed better short- or long-term management, says Julie Marquardt, the project manager on the grant. Once Minnesota enhances security measures for the Web portal, it hopes to add clinical data, such as laboratory results and diagnostic imaging. In time, it hopes to expand the record further to create a document that can be shared by all providers caring for a patient to coordinate his or her care.

Together, these projects demonstrate the challenges that providers, policymakers, and health care purchasers are likely to encounter as they implement large-scale health information technology projects in their communities. The states that implemented these projects were required to navigate complex policy, governance, and logistical issues. The challenges include how to handle medically sensitive information such as psychiatric records or HIV status; how to obtain consent from patients with diminished capacity; and how to keep providers—especially those in small, independent medical offices—engaged enough to invest in and use the technology. While many of these issues are not new to clinical practice, the pace at which they had to be resolved to meet grant deadlines compelled the states to find creative solutions. With that in mind, Quality Matters asked program leaders (and others with expertise in health information technology) what advice they would offer to those who will soon receive HITECH funds and embark on similar projects. Here's a roundup of their advice:

1. Involve stakeholders from the beginning and define those stakeholders broadly. "There's a tendency for people to think of this narrowly—equating stakeholders with physicians and hospitals," says Patricia MacTaggart, a lead research scientist and lecturer at George Washington University in Washington, D.C., who has written about state e-health initiatives. To build systems that enable providers to coordinate care, development work must include input from pharmacies, nursing homes, dentists' offices, substance abuse centers, home health agencies, and others. Managing the participation of so many will be a challenge that requires prioritization, but it is worth the effort, experts say. States that exclude such stakeholder groups from the initial discussions may re-create a problem many managed care companies had when they created "carve-outs" for management of certain categories of services such as psychiatric care, only to discover that effective case management required them to integrate those services back into the insurance product. "You have to have everyone at the table to plan broadly," MacTaggart says. For a good example of stakeholder involvement, see the case study of Alabama's "Together for Quality" program in this issue. Planners also should go out of their way to build trust among participants, because those participants are unlikely to share data essential for such projects otherwise, says Anthony Rodgers, director of Arizona's Medicaid program and Children's Health Insurance Program (CHIP). "We were very transparent," he says. "We would have meetings with all the stakeholders and show them who was accessing the data. We have an audit trail and, over time, they became comfortable that we knew what we were doing."
2. Don't surprise the end users. Make sure the intended users are included in the design stage of creating electronic health records, health information exchanges, and other technologies. "States that do it well involve stakeholders from the very beginning through to the design," says Jessica Pollak Kahn, M.P.H, the project officer for the Centers for Medicare and Medicaid Services, whose job includes overseeing and evaluating the Medicaid Transformation Grant program. While building a Web-based medical record for foster care children that combined demographic information with medical claims, pharmacy data, immunizations, and behavioral health records, Texas delayed involving providers because the process already required a great deal of coordination between multiple state agencies and the vendors who maintained these data. By the time physicians and other caregivers were invited to look at the product, it was too late to act on their recommendations. "We should have involved the providers earlier," says Yvonne Sanchez, senior health policy analyst for the Medicaid and CHIP division of the Texas Health and Human Services Commission. The state is now following up on providers' recommendations, which include adding laboratory results to the record and ensuring its interoperability with electronic medical record systems that were already in place in some physicians' offices. Mark Zuliani, the transformation grant project manager for New Mexico's Human Services Department, says it's also important to include a broad range of users in the discussions. "People talk to high-level individuals because they tend to make the decisions, but the ground level is just as important because those are the individuals that have to use it," he says. New Mexico, which used its grant to build a medical history of Medicaid patients for use by emergency department physicians, found providers used the product only sparingly—in instances when a patient was incapacitated and could not provide a medical history or when providers wanted to review medication records.
3. Investigate whether state law will permit providers to share medically sensitive information—such as HIV status or treatment for psychiatric conditions—as well as other restrictions on sharing medical data. Such laws will restrict the usefulness of the end product. In Tennessee, for example, state law required that independent clinical laboratories return results to providers and no one else. That meant providers couldn't contract with a third party—such as an electronic health record vendor—to hold and aggregate laboratory data. There were separate rules for pharmacy data. "It's really a maze you have to wander through," says Brent Antony, chief information officer for TennCare, the state's Medicaid managed care program. The rules governing medically sensitive information complicate matters further. New Mexico was required to suppress information related to HIV status, substance abuse history, and psychiatric conditions in its electronic health records. "The problem with suppression is how do you suppress it? Do you suppress the whole client?" Zuliani asks. "We suppress at a more defined level," he says, by provider type and code. But even that requires significant maintenance to ensure the system keeps up with treatment protocols and drug uses. Wisconsin changed its law to expand the amount of information that can be shared among independent institutions without the written consent of a patient. And Tennessee is investigating whether it can allow providers to "break the glass" and obtain medically sensitive information under limited circumstances. 
4. Decide whether state policy should require residents or enrollees to opt in or opt out of participation in the electronic system. In Rhode Island, residents must opt in to the health information system before their data can be added. They also may choose which providers have access to their data. Other states, such as Tennessee and Minnesota, are using opt-out strategies, which require patients to request exclusion from the system. In Tennessee, fewer than 1,000 Medicaid enrollees out of 1.2 million opted out when given the option to do so, Antony says. Both methods of gaining consent require significant investment because states must inform residents of their plans and the process for opting in or out. "It's everything from the legal small print to beneficiary outreach," Kahn says. It also requires devising a plan to reach residents with low literacy or diminished capacity. "How do they interact and give their approval for use of the data? There are many levels that need to be thought through," she says. Texas managed to sidestep the issue by building a system for children in the foster care program. Because the children are wards of the state, the state had authority to use their medical records.
5. Prepare to spend as much time helping providers implement the technology as you do designing and building it. Small providers, in particular, will need help purchasing equipment and integrating it into the workflow of their offices. Arizona helped them by creating a purchasing collaborative for electronic health records and soliciting bids for systems on behalf of physicians. The state evaluated 16 vendors in all and selected two for the physicians to consider. One product was more comprehensive than the other, but both were made available to physicians who agreed to participate in the collaborative on a subscription basis, which cost less than $1,000 per month, Rodgers says. The state plans to offer the same service to hospitals and long-term care facilities. Arizona also deployed a support team to help providers learn how to use the technology. Without having such a person available via telephone or in person, states run a high risk that providers will not fully adopt and use the technology, Rodgers says. Tennessee, which negotiates a discounted rate for networking services for its offices, extended those rates to individual providers—including those in rural communities—to ensure they could find telecommunications services and support at reasonable rates. Federally funded regional extension centers that provide technical support and a means to allow hospitals and physicians to share best practices will also help, once they become operational. 
6. To ensure interest and engagement on the part of the providers, reward them with something useful at the outset. High-demand items include current medication lists, laboratory results, discharge summaries and allergies, and a summary of patients' medical history, Rodgers says. "That's the pecking order physicians give us," Rodgers says. It's very important to do this early. "At the end of the day, if all we've added is more hassle to the doctors, they will eventually say, this wasn't worth it. Even the administrative incentives aren't worth it," Rodgers says. For more information on designing systems that enable physicians to use such technology to their advantage, listen to an audio interview with Justin Starren, M.D., Ph.D, director of Marshfield Clinic's Biomedical Informatics Research Center.
7. Find a vendor who is willing to make its methods and systems transparent, so that the state can replicate its work or expand its scope or functionality. Minnesota insisted on finding a vendor willing to do so. They "taught us what they were doing step-by-step," Marquart says. Without such a relationship, there's a risk the technology will become a black box, which only the vendor can control. Without such precautions, Medicaid programs "are all prisoners of their vendors," says Mark Frisse, M.D., M.B.A., a professor of bioinformatics at Vanderbilt University in Nashville, Tenn.
8. Consider the financial consequences of these programs on providers. "If successful, we will avoid duplication [of care]; we will avoid re-admits," MacTaggart says. And it won't take providers long to figure out they are making less money. Payment reform efforts must address this, she says. "We are all focused on getting them going, but if you don't deal with the back end, getting them going falls flat really quickly," MacTaggart says.
   

 Despite the challenges states are likely to face as they implement health information technology initiatives, Medicaid agencies are in a unique position to promote the use of health information technology and influence its impact on quality of care outcomes. Their purchasing power and the extensive relationships they have with both private payers and providers will provide an opportunity to improve coordination of care and accountability in health care organizations, both of which have been difficult to achieve.

 1. There are multiple definitions for health information technology. The term is most often used to describe computer systems and applications that allow for the acquisition, management, and retrieval of medical information for the purpose of improving health care quality, reducing medical errors, enabling disease surveillance, and improving chronic care management. Using such technology to influence the quality of care through the measurement and analysis of performance and outcome measures requires sophisticated systems, which depend on extensive data collection. Many of the projects funded by the Medicaid Transformation Grants focus on the collection of data rather than analysis of it.
2. The $46.8 billion figure includes $23.1 billion in incentives for providers in the Medicare program, $21.6 billion for providers in the Medicaid program and $2.1 billion for the administration of the program by federal and states agencies.