Hospital uses wireless technology to give doctors the edge" ... "Pharmacy vendor integrates e-business strategies in global marketplace" ... "Artificial intelligence moves to predictive modeling systems."
These are typical headlines on technology in managed care. In reality, however, progress on healthcare technology initiatives such as electronic health records, interoperability, decision support systems and e-business has come slowly. These initiatives continue to mature and improve quality and costs, but they are burdened by a lack of homogeneity and by slowly emerging standards.
Missing are articles about technologies that are advancing too fast for the businesses they serve. The reality of recent years' "progress" in integration of data into enterprise-wide data warehouses (EDWs) is that organizations are now being stifled by this "progress." More powerful and cheaper data modeling tools; data extraction, transformation and loading tools; and better server technology have all expanded the amount and types of data that an EDW can assimilate. By putting every piece of data into the EDW, end users typically don't get data in a format that they can understand and use. What once were small data silos or marts with lots of value to the end users have turned into increasingly complex EDWs that lack business value.?
Fueled by innovation in technology, increasing demands for data and the increasing IT workforce over the last five years, EDWs have reached critical mass in both size and complexity. Integrating disparate data sources across business functions to deliver the power of data mining and reporting has largely succeeded—from an IT perspective. The current trend is to use the technology available and provide an EDW that leaves nothing out. The success of implementing large EDWs has come at the cost of a significant increase in complexity and size.
But how much integration is needed? At what point does integration become so great that it limits the usefulness of the EDW? Starting in the 1990s, business units within healthcare organizations began to ask for decision support capabilities that come from timely data spanning financial and clinical boundaries. This spurned a new industry mantra—"data is king"—that drove healthcare informatics away from the multiple data silos into larger and larger data stores. Supporting this integration were developments in database design, pioneered by experts like Dr. Ralph Kimball. Technical advances in hardware, the widespread adoption of programming frameworks such as .NET, the spread of networks and the Internet have all contributed to the technical boom.
Many comparisons are made between the healthcare industry and the financial industry when it comes to the development of data warehouses. The financial industry, however, has not experienced this rise in complexity. So what is it about the healthcare industry that differs so? Is it the structure of the data itself? It would be a remarkable financial transaction that would have more variations than in medical records.
Healthcare business flaws compound the complexity of healthcare data. End users often do not prioritize their needs, so the "pie in the sky" stuff gets the same priority in the purchase or design process as the essentials. Also, too much emphasis is placed on data accuracy. A balance must be struck between completeness and practicality. An unreasonable approach rejects so many records due to data quality that the EDW loses credibility because of its inability to generate a balanced view of aggregated financial results.
Applications that bridge the complexity gap are becoming more prevalent, but they too often involve a steep learning curve, and they are not always as transparent as they should be. Layers of data marts and third-party tools now commonly sit between the EDW and the end user. Tools such as reporting suites, multi-dimensional data analysis, or online analytical processing (OLAP) are typically the only link that end users have to their data.
The managed care informatics industry is at a crossroads. Unlike during the 1990s, when IT solutions were not up to the challenges of modernizing healthcare, today's technology can create increasingly complex solutions. Looking back, it could be argued that, if the measure of success is usefulness, the adoption of the computerized patient record in the mid-'90s was generally a failure. The progress we've made in technology over the last decade has given us the power to create "successful" solutions that do not always provide value to our industry.
The future for businesses in healthcare—and healthcare informatics—lies with those who strike the appropriate balance between complex systems that work technically and simpler solutions that meet business needs. Certainly, many organizations will continue to invest at the speed of technology to capture more and more information. Ultimately, the winners will ask the question, "What should I build?" and not "What can I build?" They will develop solutions that will put the data stored in EDWs into the active hands of users in all business functions, not just the hands of developers and database analysts.
David A. Cusick is a healthcare technology consultant in the New York office of Milliman, Inc.
RETURN TO FEBRUARY 2006 TABLE OF CONTENTS
Get the latest information on Health IT and attend other valuable sessions at this two-day Summit providing healthcare leaders with educational content, insightful debate and dialogue on the future of healthcare and technology.