Data, data everywhere but not a drop of knowledge. For the past decade, businesses have been collecting data, backing it up and archiving it. Now they're scratching their heads, wondering what to do with all of it. Enterprises may be drowning in data, but much of it is sitting in silos without any effective means of sharing it or extracting value from it. Business intelligence tools are helping to transform raw data into smart information. While business intelligence isn't a new concept, it's starting to pick up steam as businesses realize they need to make better use of their mountains of data.

In previous World View reports we explored promising developments in healthcare that are made possible by clinical information technologies. Technology-related investments such as national and personal health records, picture archiving and communications systems and remote monitoring techniques offer a multitude of clinical benefits—error reduction, safety improvements, enhanced chronic disease management and improved clinical decision-making, to name a few. Fundamentally, these benefits can only be realized once a robust technology infrastructure is established and systems of work are changed to support innovation.

In our last report we highlighted how the convergence of evolutionary changes in the practice of medicine with revolutionary technologies can lead to richer data capable of improving diagnosis and therapeutics via personalised healthcare. In this report, we emphasise the importance of healthcare organisations anticipating and managing the shear volume of data produced by these existing and emerging technologies. As Ferguson points out, "the challenge today…is not just obtaining the data, but knowing how to use it." Krohn contends "the healthcare industry is awash in data and is unsurpassed in its preoccupation with gathering, storing, processing, analysing and distributing information. But healthcare data is notoriously fragmented and often incomplete, making analysis and knowledge distillation from such sources an ongoing challenge."

A growing number of technologies for integrating and performing structured analyses of data from disparate sources are competing to win the day for healthcare organisations. This report provides an overview of several of these approaches within the arena of business intelligence—the application of computer technology that helps organisations to adopt best practices—which has emerged in healthcare as an important means of informing cohesive strategic and clinical decision-making.

Aids to Navigation: Business Intelligence Tools

Business intelligence is a broad concept encompassing the processes, software and technologies for gathering, storing, analysing, and providing access to data to help users make better decisions. In essence, it is "the art of sieving through large amounts of data, extracting information and turning that information into actionable knowledge." A given business intelligence system exploits numerous tools in order to meet specific goals, with progress measured using key performance indicators against best practice benchmarks.

As a store of static summary data from multiple internal and external sources, a data warehouse is a database repository providing ready access to population and health systems data that has been integrated from across the organisation (analogous to a library). A data mart similarly consolidates data, but only for a specified program area or reporting purpose (analogous to a section in a library). These repositories are fed by operational systems, such as an electronic patient record system (providing up-to-date data on individual clients) or a clinical data repository of real-time operational data about groups of patients often within a particular healthcare programme.

Although initially intended to support financially oriented analytical applications, data warehouses are increasingly being integrated with clinical systems and operational applications. A high level of data integration from various sources facilitates advanced data analytics and reporting in order to leverage the full potential of business intelligence. This is achieved using data extraction and analysis tools such as data mining (analysing large databases of complex data to discover previously unrecognised patterns and useful trends) and OLAP-online analytical processing (similar to data mining but with the added capability to analyse data in multiple dimensions for trend analysis and forecasting). Users control these processes using sophisticated computerised dashboards for viewing customised information.

The class of applications that enable these processes is called "intelligent software" or "advanced decisioning systems." These and other business intelligence technologies have been used for years in healthcare for financial and administrative purposes. Now business intelligence is also helping healthcare organisations with diagnosing and treating patients with long term conditions and evaluating alternative treatments based on outcomes analyses. For example, data from operation-level clinical systems can be used to prioritise heart bypass surgery versus angioplasty by looking at how quickly patients are discharged and whether they're readmitted.

Additional capabilities include the ability to distinguish previously unrecognised disease patterns, identify at-risk patients, and review the performance of individual physicians. As Wyatt points out, the need for operational performance monitoring, e.g., through the use of scorecards as measurement tools, is driving health care organisations to "take a strategic approach to measuring performance across the entire organisation and…to feed that information into their planning activities to more accurately predict operational factors and react faster to problems. The result is an improved ability to reduce costs and better balance quality care with operational expenses."

Furthermore, thanks to ever improving and easy to use database and interface designs underpinning business intelligence tools, the power of analysis now lies in the hands of healthcare professionals for daily tactical decision-making, not only with management for occasional strategic decisions impacting healthcare delivery. Access to these capabilities on the front-line energises a working culture that supports positive change. According to Sanders, the objective of data warehousing is measurement in order to gain understanding that informs behavioral change and continuous quality improvement. By extension, business intelligence is predicted to change the healthcare landscape through enhanced efficiencies, improved outcomes and better decision-making. Many examples already demonstrate this new reality.

Desirable Destinations: Healthcare Success Stories

In Canada, the Vancouver Hospital and Health Services Centre has, until recently, had little means to analyse its volumes of patient and operational data generated from over 37,000 acute-care admissions and 81,000 emergency visits annually. Using a powerful business intelligence tool, the multidisciplinary Quality Utilisation Information Support Team (QUIST) was able to quickly build subject-specific data marts (or PowerCubes) to measure performance and communicate indicators across the organisation. Even though the current system so far contains only 10% of the Centre's data, hundreds of users are finding 80% of the information they need to make better, informed decisions. A valuable spin-off from the ease of deployment and use of the tools has been enhanced interest in business intelligence and increased demand for support of better decision-making across the Centre. This finding is particularly important because, according to Sanders, "the technology behind data warehousing and analytics is useless without the personal and cultural commitment to use it."

The stories are similar in the United States. Intermountain Healthcare (IHC), a community-owned, nonprofit healthcare system, provides health insurance and medical services to over a million Utah and Idaho residents through its 22 hospitals and 80 outpatient clinics. IHC has been ranked the top integrated health system in the US five times in the last six years. IHC's award-winning technology systems include a mature very large data warehouse containing over 25 years of medical records. This system enables IHC to conduct research and analysis in order to improve patient care and lower costs. In fact, by using their computer system to identify best clinical practices among its physicians, IHC is saving the lives of as many as 1,200 cardiovascular patients each year. As Sanders reports, business intelligence enables analysts to now spend 90% of their time—which they used to spend collecting data—working with clinicians and managers to improve their processes.

A BayCare hospital in Florida has managed to reduce the time it takes to diagnose and process a heart patient by 20 minutes, thanks to computer systems that enable clinicians to analyse procedures for treating suspected heart attack victims. In Cincinnati, the Children's Hospital Medical Center has reduced medication errors by half by analysing medication orders and feeding the results back to a medication-administration system. The St. John Medical Center in Tulsa, Oklahoma has reduced the number of transfusions leading to negative reactions by 18% and reduced the number of transfusions performed by 22 %—a savings of $1.4 million annually. The Center's data warehouse is now being used to monitor bacteriology test-order patterns to identify a possible terrorist attack. As Dr. Terry Dolan, St. John's President asserts, "We feel we haven't even scratched the surface of what data warehousing can do for healthcare." Business intelligence capabilities will continue to advance as a key infrastructure component and enabler as developers tackle the ongoing challenges of data integration and change management.

Hazards to Navigation: Challenges to Integrating Data

According to a PR Newswire report, data integration is expected to emerge as a key underpinning of the U.S. Department of Health and Human Services' National Health Information Infrastructure. Successful integration, however, depends on being able to get "the right data into the right hands at the right time to enable health system decision-making" (the vision for health information management in British Columbia, Canada). Krohn describes the challenge as "To be useful, business intelligence must provide access to actionable information, but not too much information, and only the information that a user genuinely requires. The problem is double-edged—too little business information shared inhibits decision making, but too much business information produces layers of analysis, process, approval and, ultimately, inertia."

In fact, as Zimmer contends, "paralysis by analysis" is just one of the reasons that data warehouse projects fail. Other impediments to success include inadequate training, struggle for control between IT and business users, insufficient short-term (in addition to long-term) goals, and a poor understanding of the data required by users and the technology itself in order to optimise results. Disparate user information needs and existing (paper and electronic) source systems where the data reside represent core challenges to data integration. As of March, 2004, according to Whiting, less than 15% of all clinical data was digital. Much of the data that is digitised is formatted for billing purposes, not clinical analysis. Healthcare organisations are overcoming these barriers by using business intelligence tools that are "integrated into operational applications and designed to work with data they produce."

By focusing on standards for capturing, storing, extracting and reporting clean healthcare data in the context of well-articulated user needs, a healthcare organisation can move closer to becoming a ‘zero latency enterprise.' Originated by the Gartner Group, this term refers to computer systems that enable an enterprise to react more swiftly to changing conditions and business issues. As Krohn summarizes it, "The term information latency refers to the amount of time it takes for data to take effect between different parts of a computing infrastructure. Through application and data integration, the zero latency enterprise provides an enterprise-wide capability to immediately see and act on business intelligence. At full maturity, the zero latency enterprise has removed all delay from its operations so that business events that occur anywhere in the organisation can immediately trigger appropriate actions." Once the challenges of data integration are sufficiently met, organisations will be able to move beyond the analysis of static data in order to predict the future with the help of real-time technologies.

The Wave of the Future: Predictive Analysis

The future for healthcare is promising. As the thirst for up-to-the minute information and knowledge-building gains momentum, business intelligence technologies such as real-time data warehousing, predictive modeling and pattern analysis are emerging to meet the demand. Himmelsbach reports that the business intelligence market grew 10 per cent in Canada from 2003 to 2004 with a predicted compound annual growth rate of 8.1 per cent over the next five years. Data analysis is moving from retrospection to prediction and the benefits are already clear—more efficient operations, reduced costs, increased quality and "a healthier bottom line," which in healthcare also translates to saved lives.

References

Akhtar M, Dunn K and Smith J. Commercial Clinical Data Warehouses: From Wave of the Past to the State of the Art. Journal of Healthcare Information Management. Vol. 19, No. 2, Spring, 2005:20-26.

Allmon A. Advanced Analytics: A Powerful Secret Weapon Against Health Care Fraud. HealthCare Review. July 19, 2005.

Barlow S. Intermountain Health Care Integrates Data Warehouse Information with Data Junction. DM Review. December 2001.

Cognos—Customer Successes. Vancouver Hospital and Health Services Centre. 2005.

Ferguson G, Mathur S and Shah B. Beyond Data Warehousing: New Applications Turn Information Into Profitable Insights. Accenture. Outlook: Point of View. May 2005.

Gurevich H. Surviving the Enterprise Integration War. Health Management Technology. February 2004.

Jennings P, Project Director of HNData / BC Project, Ministries of Health. Data Warehouse Information Utilities and Health Sector Analyses. Presentation to the School of Health Information Science, University of Victoria, Canada. July 17, 2002.

Koh H and Tan G. Data Mining Applications in Healthcare. Journal of Healthcare Information Management. Vol. 19, No. 2, Spring, 2005:64-72.

Krohn R. Data Analytics Throughout the Healthcare Enterprise. Journal of Healthcare Information Management. Vol. 18, No. 2, Spring, 2004:15-18.

Krohn R. Quick Study: Healthcare business Intelligence and Real-time Decision Support Systems. Journal of Healthcare Information Management. Vol. 18, No. 3, Summer, 2004:14-16.

PR Newswire. HIMSS Advisory: Ascential Software to Highlight Business Benefits of Enterprise

Data Integration to Healthcare Payors and Providers. February 9, 2005.

Sanders D, Intermountain Health Care. Designing, Developing, and Supporting an Enterprise Data Warehouse (EDW) in Healthcare. 2002.

Sanders D. Business Intelligence, Data Warehousing, and Software Safety in Healthcare: Two Heroes and a Villain. Presentation: February 2005.

Tyshynski B. Data Warehousing as a Strategy to Managing Information Resources: BC Health Data Warehouse Experience. Presentation to the School of Health Information Science, University of Victoria, Canada.

Whiting R. Analytics Move to The Clinic. Information Week. March 1, 2004.

Whiting R. Data Analysis to Health Care's Rescue. Information Week. September 24, 2001.

Wikipedia. The Free Encyclopedia.

Wyatt J. Scorecards, Dashboards, and KPIs Keys to Integrated Performance Measurement. Healthcare Financial Management. February 1, 2004.

Zimmer H. Data Warehousing: Are You on a Path to Success or Failure? Presentation. The Data Warehousing Institute, 1998.