As patient care organizations nationwide prepare to report on meaningful use quality measures, those with enterprise data warehouses may find they have a head start. The more advanced among them are establishing workgroups to create dashboards to analyze progress and identify gaps that need addressing.
The University of North Carolina (UNC) Healthcare System in Chapel Hill, N.C., has been working on its Carolina Data Warehouse for Health for almost four years. And Donald Spencer, M.D., UNC's associate director of medical informatics, now sees the project as a necessity for meaningful use quality measure reporting.
“In an institution as big as ours, there is not one comprehensive program you can turn to in order to get a clear picture of what is going on,” Spencer says. Like many health systems, UNC uses different software vendors for its core electronic health record (EHR), lab system, radiology information system, and emergency department IS. “Getting data from all these diverse systems for hospital quality improvement reporting would be very laborious without a single data repository,” he adds.
BUSINESS INTELLIGENCE GIVES PROVIDERS THE ABILITY TO USE DATA TO BETTER UNDERSTAND PATTERNS RELATED TO QUALITY AND PATIENT SAFETY AND IMPROVE OUTCOMES.-LORIN BIRD
“Now, in a smaller system, if the vendor has a clinical warehouse where EMR data can be copied and put in a searchable database, you could get most of this information,” he stresses. “But I like that ours is more than just clinical data; we also have administrative data and ICD-9 diagnosis codes.”
Spencer is not alone in seeing back-end data warehouses and front-end business intelligence (BI) reporting tools as crucial to meaningful use efforts. In a 2010 survey of health system executives undertaken by the Orem, Utah-based KLAS Research, 69 percent of respondents said BI solutions would play an important or critical role in meaningful use. One provider told KLAS, “Never before have we been so carefully scrutinized, and the BI system helps us provide the information and reports needed to be successful.”
KLAS research director Lorin Bird says launching a BI effort, while important for meaningful use reporting, is also about creating a vision of improved performance and technology-enabled healthcare. “BI gives providers the ability to use data to better understand patterns related to quality and patient safety and improve outcomes,” he says. It can also help them marry financial and clinical outcomes data. With BI tools, they can prove that better clinical outcomes lead to a reduction in cost, and that more attention to chronic care can lead to savings, he notes.
Bird says very few U.S. health systems have sophisticated enterprise data warehouses that pull data from both clinical and financial systems and offer end-user reporting tools. “I would say it is somewhere between 50 and 100,” he says. But he projects that number will double in the next two years. “Based on how many people are telling us they plan to bring together clinical and financial data, I think if there are 100 now, in two years there will be 200. If they have a BI solution looking at general financials, by then they will be looking at the other side of the house.”
BI Tools in the Community Hospital Setting
An enterprise data warehouse may be a daunting project, especially for community hospitals. Paul Alcala, vice president and CIO of NorthBay Healthcare, a two-hospital system with fewer than 200 beds based in Fairfield, Calif., says he has seen hospitals find out the hard way that data warehouse projects are more complicated, time-consuming, and expensive than originally anticipated. “User expectations exceed their ability to deliver and they go over budget,” he says, “and a lot of projects fail.”
Alcala decided to start smaller with data marts for specific internal audiences. Using BI software from Dimensional Insight, Burlington, Mass., Alcala believes he can accomplish about 85 percent of what a data warehouse could do for about 50 percent of the cost. Launched six months ago, the NorthBay project started with several dashboards for revenue analysis. “We will create a dashboard around meaningful use to assess where we are,” Alcala says. “We think we are in pretty good shape, but we do have some gaps. Stage 1 is just the beginning,” he notes. “Meaningful use will make the use of BI grow like crazy.”
Alcala says that he had to start slowly because a culture of data analysis really didn't exist at NorthBay. “We are working on creating a data-driven culture, but you can't just snap your fingers. It takes time,” he adds. “We are six months into what I see as a three-year journey to get up to speed in culture, tools, and analysis.”
GOVERNANCE IS KEY
Developing a governance structure was one of the first orders of business at the Carolina Data Warehouse for Health, Spencer explains. “The priority is meaningful use for a large part of the organization, but if you just got a multimillion-dollar research grant, you have a completely different priority, so we needed the governance structure to prioritize,” he says.
At UNC, an operational committee spins out workgroups on particular topics. For instance, a clinical group initially looking at diabetes turned into the ambulatory quality improvement workgroup, which has now morphed into the meaningful use workgroup. That 10-person multidisciplinary team, which includes a pharmacist, a BI analyst, clinical process improvement people, and several other clinicians, is meeting once a week to create a dashboard to analyze where they are on meaningful use measures and how they can improve on them. “For instance, if they need 40 percent of prescriptions to be e-prescribed, this group finds out where in the data warehouse to get that information and how much e-prescribing versus printing is done now, so we have the numerator and denominator,” Spencer says.
Sue Schade, vice president and CIO of the 777-bed Brigham and Women's Hospital (BWH) in Boston, says her organization has a team devoted to the reporting requirements for meaningful use. It will include data pulled from the BWH data warehouse, as well as ambulatory EHR data from the quality data warehouse at its parent organization, the Partners HealthCare integrated health system. BWH is seen as a pioneer in the use of business intelligence, with a “balanced scorecard” effort that feeds key performance indicators to clinical and operational leaders in order to improve patient care quality and efficiency.
Schade says the hospital's several years of experience with BI reporting should make the meaningful use reporting easier. “The data warehouse is the foundation for all our analytics,” she says. “If you haven't dealt with a warehouse and reporting issues, I think meaningful use will be more difficult.”
Collaborative Data Warehousing
Medical groups and hospitals that lack their own data warehouses may be able to take advantage of associations’ efforts at collaboration. For instance, the Alexandria, Va.-based American Medical Group Association recently launched a project called Anceta that features a data warehouse to offer medical groups a detailed comparative view on different subgroups of their patient populations.
One AMGA goal is to identify unwanted variation in care processes, clinical outcomes, or standardized cost. “It allows medical groups to collaborate and share best practices in a way that is more data-driven and scalable,” says John Cuddeback, M.D., Ph.D., AMGA's chief medical informatics officer. “We can put data in their hands with a visually oriented analytical tool that isn't any harder to use than the Travelocity Web site.”
AMGA's technology partner, the Boston-based Humedica, has a tool to extract data from transactional systems. “If we expected the medical groups to do the mapping themselves in order for us to get apples-to-apples comparisons, it would never get done,” Cuddeback says. Humedica goes into native tables of EHRs, practice management, and other hospital systems to extract data.
Launched in March 2010, Anceta now has 11 medical groups participating and 8 million patients in its database, with collaborative projects around diabetes and congestive heart failure in the works.
Another example is the Quality Data Center run by the Massachusetts eHealth Collaborative. As part of a health information exchange pilot project several years ago, MAeHC built a data warehouse for evaluation purposes. Once the pilot project ended, other health players began asking MAeHC for help, says President and CEO Micky Tripathi. For instance, 621-bed Beth Israel Deaconess Medical Center in Boston, which is managing the implementation of two EHR platforms for affiliated physicians, has asked MAeHC to extract data on 1,200 doctors, so Beth Israel can do internal benchmarking and report externally, including to insurers’ alternative quality contracts, meaningful use and the Physician Quality Reporting Initiative (PQRI). MAeHC is also providing the data warehouse for a patient-centered medical home project with 300 physicians in the Albany, N.Y., area.
“We are helping physician groups with EHR implementations and it is not just plug and play for reporting,” Tripathi says. “The chance that you will get the data you need out of that one EHR system is close to zero.”
He adds that a solution like the Quality Data Center may buffer its participating practices from all the different reporting regimes and changes in reporting requirements. “If they send us a clinical superset of what they need to report on, we can shield them from the logistical issues around reporting as long as the data is there.”
Her suggestion to CIOs still in the exploratory phase is to start small and get early wins with groups of employees already trying to do measurement for performance improvement.
She also suggests limiting how many things you measure, and making sure they have value, but she adds that meaningful use quality measures will help organizations, because it makes pretty clear which things to prioritize.
A MEANINGFUL USE DASHBOARD
One former health system CIO saw such an opportunity for data warehousing in healthcare that he left his position last year to join a startup in the field.
Herb Smaltz, the former CIO at the 872-bed Ohio State University Medical Center (OSUMC), is now CEO of the Columbus-Ohio-based Health Care DataWorks, a software venture that spun out of Ohio State two years ago and has developed a meaningful use dashboard for health systems. Health Care DataWorks, which claims to offer an “enterprise data warehouse in a box” for healthcare provider organizations, says its meaningful use dashboard provides a single location where an organization could see how it is performing on meaningful use indicators.
Most core health IT vendors will offer some way of reporting out of the EMR, Smaltz says, but hospitals in a best-of-breed environment may have difficulty pulling that information together in one place, he warns. Without an enterprise data warehouse, you may have trouble beyond the first stage of meaningful use, he says, because “as we move to new reimbursement mechanisms and as more interoperability is required between community providers, you need to pull data from a patchwork quilt of sources.”
In the first year of meaningful use requirements, hospitals may be able to get that data manually or through data marts they already have, he says, but “as the bar is raised you won't be able to avoid a data warehouse, the same way some tried to avoid getting an EMR but most realized they just couldn't avoid it.”
Smaltz acknowledges that the transition to an enterprise data warehouse can be difficult. He recalls a recent consultation with a large health system in the Southeast. “They had five different data marts with five internal owners and five different software vendors,” he says. “There are lots of issues to sort out regarding data governance if they are going to move to one enterprise data warehouse. It is not an insignificant task.”
The Moffitt Cancer Center & Research Institute in Tampa, Fla., has had a data warehouse for several years as part of its Total Cancer Care project, which gathers demographic and clinical data on nearly 40,000 patients. It is used largely for research but also for quality reporting and clinical and financial analyses.
But Mark Hulse, R.N., who has been vice president of information technology and CIO at Moffitt for less than two years, says the data warehouse is already being redesigned. The previous iteration, developed in-house, was designed to make it flexible to get data into it, so if transactional systems changed, it wouldn't be a problem, but it proves complicated to get data out of it, he says.
The new version will use a hub-and-spoke model. “The core information will be in the data warehouse, but if departments want to collect more specialized data, they can keep that in their own departmental databases,” he explains.
Researchers would like the new data warehouse to have the capability to capture robust metadata about the data source and versioning information. “It is important to know which version of a patient questionnaire a patient filled out, especially for published research,” Hulse explains.
Healthcare Informatics 2011 January;28(1):18-23