|
|
For other articles and previous issues click here. August 1, 2005 Inside
Healthcare’s Warehouse District It’s not exactly trendy, but it is worth taking a stroll down data warehousing’s aisles to see whether there is something on the shelves to make your healthcare organization run more efficiently. In today’s information-driven society, those with the most accurate data often have an advantage—whether in healthcare or any other industry. Data warehousing in the healthcare field provides both clinical and financial data so administrators can use them as a way to seamlessly improve internal processes. In case of disaster, warehousing should be able to provide high-speed, instant access to back-up data. Perhaps the most critical function of a data warehouse is its ability to copy transaction data specifically structured for querying and reporting. That is one reason why many information technology (IT) managers working in the healthcare field admit that in the future every hospital or healthcare system will need a data warehouse to stay competitive. Indeed, information management is a high priority among hospitals and health systems. The ability to draw data from reliable and immediate internal sources provides hospitals with many potential rewards, including higher customer satisfaction, efficiency, and profitability. In theory, data warehousing can help hospitals provide solutions and services for knowledge creation and discovery, decision support, patient management, quality, and disease management to meet their strategic goals. Despite the benefits, some hospitals find that setting up such a warehouse is cost prohibitive or too time-consuming based on the resources (both financial and manpower) available. Most healthcare IT experts realize that the short-term cost will be outweighed by the long-term benefits but the money simply may not be there. Mike Smith, director of application services at the University of Mississippi Medical Center (UMC), does not utilize a true data warehousing system but instead uses data mining as a way to provide administrators and managers with important clinical and financial data. Smith and his staff are progressing toward what they hope will some day be a true warehouse. In the meantime, he recognizes the advantages it will provide. “A warehouse is designed for upper management to conduct ‘what if’ scenarios,” he says. “Data mining is not an IT solution and a warehouse is.” That’s because data mining requires significant in-house training to allow internal staff to manipulate the data as needed. And while having a successful data mining process is much preferable to a more cumbersome data process, Smith and UMC administrators understand a warehouse does provide a significant return on investment. “There needs to be internal analysis first,” says Smith, who indicated that UMC is currently undertaking such an analysis. “We have tremendous support from our administration to move forward in looking into a warehouse solution.” Data warehousing allows Jacksonville, Fla.-based Nemours Health System, one of the nation’s largest pediatric subspecialty group practices, to pull together massive amounts of medical information in a short amount of time, according to Ed Todd, data warehouse manager for Nemours and Alfred I. duPont Hospital for Children. Todd, who previously worked with data warehousing in the financial services sector, says the technology also provides a source of integration between financial and clinical data. “There is tremendous consistency with the amount and types of data you can pull and the data can be accessed in a timely manner,” he says. Nemours’ data warehousing capability was put to good use when the FDA made a series of recalls over the past 18 months. Each time a drug was recalled, Todd and his staff were able to produce a list of all Nemours patients who had taken that medication, sort the list based on physician, and get in touch with those physician offices so the appropriate steps may be taken. During each recall, it takes less than one hour for Todd to pull that amount of data. In fact, that’s generally regarded as one of data warehousing’s best features—it provides hospitals with the ability to look at a large number of cases simultaneously so patterns can be identified. That can be a patient benefit some hospitals may choose to market. “There is a real quality-of-care aspect to data warehousing that patients may not realize, but it’s there because of the medical analysis you can conduct,” says Todd. He says hospitals and systems can use this IT function to develop a “best practices” process by which antibiotic effectiveness, symptoms, and other patterns can be recognized. This warehouse of information has helped the Dana-Farber Cancer Institute schedule patient appointments more efficiently. As first reported in an April 2003 article in Information Week, the teaching facility for Harvard Medical School began regularly downloading data from patient scheduling, registration, and billing systems into an Oracle data warehouse. These reports then identify the best times to schedule patient visits. The data warehouse is also used to better plan staffing and facility needs and contains clinical information, including data from an outside company that provides laboratory services and information from Dana-Farber’s pharmacy system. Even the marketing department has benefited from Dana-Farber’s data warehouse. It has become a main source of information used in patient mailings about new treatment options, educational seminars, and changes in procedures. “This approach of centralizing responsibilities takes much of the burden off individual departments,” says John Orechia, clinical reporting team leader. The Mountain States Health Alliance in Johnson City, Tenn., has employed an in-house data warehouse since 1996 and used it to generate reports for departments across the system. The alliance acquired five area hospitals in 1999 and the data warehouse was used to help seamlessly integrate the data. More recently, Mountain States has used its warehouse in conjunction with Internet-based technologies to quickly integrate data from multiple sources and locations, according to Lynn Northcutt, manager of application and development integration. “It’s the only place where you can get enterprisewide reporting while getting the same answers,” she says. Todd says data warehousing has allowed Nemours to move forward with a telemedicine project through which children in rural areas can use broadband communication tools to receive monitoring and follow-up at home without having to go to the hospital. “[Through data warehousing,] we will see more retrospective studies on data for diseases and conditions,” he says. “For example, it allows us to monitor emergency room data so we can pinpoint a potential biological outbreak.” Northcutt says her department has set up daily maintenance feeds and quality audits in conjunction with the data warehouse that are completely automated. From a training perspective, users can access an internal intranet for a data dictionary, interface information, and audit and internally check guidelines to help them better use warehouse data. Toronto-based IT consulting firm Infosys Technologies Ltd. has helped several healthcare clients reengineer their data warehouses. It has improved warehousing capabilities by adding a modular, scalable architecture to reduce batch time. On occasion, Infosys maintains client data warehouses, which helps free the client’s internal resources to tackle other pressing tasks while lending flexibility in operations, according to Infosys media spokeswoman Tina George. The cost of implementing data warehousing can be significant, but once it has been developed, organizations may view it as a critical investment. Todd estimates that Nemours’ data warehousing comprises no more than 0.5% of the system’s total expenditures. That includes the costs associated with using “off-the-shelf,” brand-name IT equipment such as IBM servers and an Oracle database. The financial aspect of developing a data warehouse is a continual challenge for UMMC’s Smith. He says while data mining can provide administrators with important trend statistics and profit/loss reports, it requires more significant in-house training than a warehouse, where an IT team or outside vendor analyzes the data and provides it to administrators and managers already manipulated for their needs. “We try to insulate the end user from as much work as possible with data mining, but again it’s not an IT solution,” says Smith. There are plenty of challenges to contend with when managing such a data warehousing project. Todd says the more successful the project has been at Nemours, the more demand has been placed on his department’s broad shoulders. “The need for data is always escalating,” he says. As of June, Nemours had amassed nearly 1,000 gigabytes of data. Another challenge is hiring qualified IT personnel who deal well with others and have either a basic understanding of the healthcare field or are eminently trainable. “It takes a pretty broad and unique skill set to find competent staff who are comfortable working in this environment,” Todd notes. Perhaps the biggest challenges occur when an existing
database needs to be reshaped by an outside vendor to become a true
data warehouse. George refers to a health benefits client that had
inefficient extract, transformation, and loading mappings, mismanaged
code that was of a lower standard, and a bulky, slow application
because of the vast amounts of data. “A significant overhaul
of an existing database can really become a headache,” admits
Todd. “There is a trust and relationship-building mechanism involved with handling this type of data and that trust comes more readily when it’s an internal function,” says Todd. “An outside vendor might slightly manipulate the data in some way and that could lead to problems.” Yet for those hospitals that prefer to outsource much or all of the work involved with establishing a data warehouse, there are plenty of options. Netezza Corporation’s Netezza Performance Server architecturally integrates servers, storage, and a relational database into a single, enterprise-class appliance created from the outset to process queries faster and enable complex analysis not previously possible. This comprehensive analysis enables healthcare clients to harness information located in tens, even hundreds, of terabytes of data to lower costs, reduce risk and fraud, strengthen customer relationships, and keep pace with ever-changing market trends. “Healthcare is one of the most data-driven and intensive industries in the world, making informatics a high-return, high-impact solution for forward-thinking organizations,” says Jit Saxena, cofounder and CEO of Netezza. “Data warehousing can help hospitals make faster and better business decisions while providing more strategic services to their customers.” Saxena says the goal is to provide customers with a warehouse of data that provides a dramatic increase in flexibility, speed, and overall performance along with a lower total cost of ownership. He contends that outsourcing this work is hassle-free without any worries about privacy or other issues. It becomes a hospital preference, he adds. “The informatics business intelligence products that we provide are some of the most valuable resources a hospital can have today, so it’s an important decision and hospitals need to determine if it makes more sense for them to handle it in-house or to outsource the work. It’s a control issue as much as anything,” says Saxena. Smith says when UMC does go forward with a true warehouse, it will do so by researching the services offered by various vendors. Because UMC is a state-owned institution, Smith will go through a request for proposal process to determine the most appropriate vendor based on qualifications, price, and more. “I think there are a number of warehouse solutions on the market today that would work for us,” he says. Smith says the bottom line is to meet the needs of UMC patients while improving the internal processes used to provide the hospital with important financial and clinical data. He admits the time will come when nearly every healthcare system and hospital has a true data warehouse. But for now, these institutions can only do what they can afford and many IT departments must make do with limited resources and a less-than-optimal data solution. “Our goal is always to maximize the amount and types of services we offer to our patients,” he says. “What we have got gives us a lot of benefits but the mining process is nothing more than a stopgap measure.” Northcutt agrees that more healthcare systems and hospitals are adding and updating their data warehouses and asserts that having one is required based on the direction of the healthcare industry. “More and more we utilize integrated delivery with multiple hospitals and in that case having a warehouse to seamlessly share and query information is critical,” she says. “And as we partner more with other institutions, the importance will only increase.” — Mike Scott is a freelance writer who has contributed to more than 70 magazines, newspapers, and Web sites on numerous topics—from business to healthcare to technology. He lives in Waterford, Mich.
— MS |
 |
 |