One of the fascinating paradoxes of our national health care system is that although hospitals and health care centres have traditionally gathered enormous quantities of data, there is still a plethora of unanswered questions about what’s working well in health care and how we might improve its efficiency and effectiveness. Although we have a great deal of data, we never seem to have great data – the kind that helps us understand what we’re doing and where we’re going.
Ironically, it is this vast bulk of acquired data with its divergent resident locations and task-specific views that makes finding the right information problematic for employees, professionals, managers and executives. Fortunately, at the same time that standardized metrics are beginning to evolve within hospitals, the methodologies and technology tools are now at hand to retrieve and manage intelligible data efficiently. This bodes well for improving Canadian health care.
One of the driving forces behind the application of performance indicators within hospitals is The Canadian Council on Health Services Accreditation, a non-profit, non-government organization that helps health service organizations across Canada examine and improve the quality of the care and services they provide to their clients.
The Council provides the opportunity for voluntary participation in an accreditation program based on national standards. The value of accreditation is derived from the internal self-assessment that an organization undergoes in preparation for a survey visit and in the consultative peer review process that is part of each visit. All teaching hospitals in Canada must be accredited.
In 1995, the Council introduced its Client-centred Accreditation Program (CCAP), an initiative that explores how an organization delivers care and service (processes) and how it knows it is performing well (outcomes). One of the requirements of CCAP is that organizations identify and monitor performance indicators as part of their efforts to improve the quality of their care and service.
During their 1997 accreditation on-site survey visit, Council representatives recommended that Vancouver Hospital and Health Services Centre (VHHSC) take a consistent approach to identifying and monitoring performance indicators that are meaningful to them and their clients for improving health care and service. (The Council does not yet have a list of mandatory or recommended performance indicators. It is currently most interested in how teams use and refine performance indicators that are significant to them.)
Shortly thereafter, the hospital board established a mandate for a management information system to monitor and evaluate the effectiveness and efficiency of its programs and services within a framework of accountability. At that time, the hospital’s information systems department and its resources were focused on Y2K compliance issues and the installation of the hospital’s patient care management system. Therefore a separate task force, the Quality Utilization Information Support Team (QUIST), was established to develop a system that satisfied both the Council’s recommendations and the board’s requirements.
“I had about three months to pull the necessary indicator information together and build what we needed with the three other members of my team,” remarks Sue Emmons, Manager of QUIST.
With formal training in epidemiology, extensive decision support systems expertise, and 25 years’ experience within the health care sector, eight of which were in the administration of records and health communication, Emmons seemed the ideal candidate to help health care managers identify their information needs.
The other members of the interdisciplinary QUIST development team included an industrial engineer with significant database design experience, a quality management expert and an information sciences expert with hardware, software and networking experience.
“Very quickly at the onset of the project we had to choose whether to proceed strategically into a data warehouse environment or tactically into a data mart situation,” Emmons notes.
WAREHOUSE OR DATA MARTS?
Emmons and her team investigated six different vendors’ data management solutions, four of which were fully integrated enterprise-wide data warehousing solutions carrying price tags ranging between $250,000 and $500,000. Three of these included sizable quotes for additional consulting to ensure that a standard approach would be implemented and that data quality would be properly controlled from the beginning.
“Because of the tight timeline we had in which to build what we needed, we chose the tactical route,” Emmons explains. “We had also established a few criteria that were absolutely mandatory. Whatever solution we selected had to be web deployable. I wanted every user to have the same consistent view. Our solution also had to be easy to install and up and running in a short time.”
Emmons believed that the best-of-breed product in the category was the Business Intelligence solution for small-to-midsize enterprises from Cognos. The tactical approach involved using Cognos PowerPlay software to develop a series of departmental data marts that would contain subject-specific data structured for a particular function or type of analysis. These ‘cubes’ could then ultimately evolve into part of an overall enterprise solution, satisfying the immediate needs of the business area, while meeting the long-term strategic goals of a fully integrated data warehouse. The software takes one to two hours to install, and a cube can easily be developed within three to four hours.
Starting with the executive team and director level of the institution, Emmons conducted user needs assessments to help identify data requirements and technology solutions, and discussed with management what approach should be applied to developing performance indicators.
“Our executive team decided to use a balanced scorecard approach and needed to be able to deploy indicator information across the organization. Before, this diverse information was coming from all over the place. We wanted to be able to pull up-to-date information all together and make it accessible from one central source in a standard format so there would be only one version of the truth.”
BUILDING THE DATA MARTS
Using Microsoft Front Page software, QUIST constructed a home page, featuring eight button links, including an indicator panel. Indicator information, which is used to measure the quality of the hospital’s use of its resources, is the result of extensive data collection from all over the institution. Cubes involved are deployed from the server with data coming in to the cubes from multiple sources. Users enter the application through the homepage and click on button links to access an individual cube.
PowerPlay Web software enables hundreds of users to transparently retrieve data from a wide variety of different databases at separate locations. Yet, from the user’s perspective all of the information on the homepage appears to be integrated because it resides in one convenient place. From a data management perspective, the software enables users to access cubes centrally and maintain cubes from one point of control for all users. Although users can manipulate the data for their own use, no one can actually alter it or damage it. Only QUIST can change or update it. The team is also responsible for staging the data, making whatever modifications are necessary to ensure its accuracy and suitability for user access.
“What we’ve attempted to do is pull onto this one homepage the management information that managers need,” Emmons stresses. “We may extract only ten percent of the total volume of information resident in an institution of this size but it’s meeting 80 percent of the management needs. It’s a great way of maximizing our resources. “
Another component of the home page is an organization chart that allows users to point and click at the hospital’s various departmental areas and view a balanced scorecard reflecting the goals for each of them. By pointing and clicking on any indicator on the scorecard list, the user is provided with an expanded view, complete with a full definition of the indicator and a performance graph of actuals for that indicator with targets, benchmarks and action points highlighted in colour. When users access these pages to review their targets, they’re expected to complete the comments for each of their action points. Yet to be truly effective as a change management tool, a good change management approach and senior management buy-in must drive the process.
To date, QUIST has constructed ten cubes encompassing workload, payroll, patient satisfaction, demographics, operating room, four-year trend data and a variety of other specialized performance indicator groupings. The project team defines its most urgent applications from regular user needs assessments. By asking users to identify their top three information requirements, common threads are identified that lead QUIST to its next data cube development task.
What’s truly exciting is that answers for improving resource use are already beginning to materialize. For example, using a cube designed for operating rooms, the hospital was recently able to monitor and assess the impact of a doctors’ action in British Columbia. By determining which patients were delayed and for what treatments, VHHSC was better able to allocate remaining resources to priority patients.
What’s more, response time for many types of queries has been significantly shortened. In the past, completing a query on a specific treatment (e.g. vascular surgery) being performed at the institution could take almost a day. Now, the same information search can be achieved in about three minutes. And with the ability to use multiple levels of stratification, power users can achieve outstanding response time and accurate results to highly complex questions.
Once a user has created a report using PowerPlay Web it can be bookmarked and indexed on the homepage, so if there’s a request that several people need, the query can be completed and saved for future use by others.
“This is not ‘hot’ on-line real-time information, like one uses in patient care. This is cold, archival information,” Emmons emphasizes. “You’re looking for trends and patterns over a period of time that may help you manage resources better in the future.”
BIG POTENTIAL PAYOFFS
As budgets become more restrained, whether it’s through inflation or the need to introduce new programs, the system facilitates the ability to identify areas where improved utilization might be possible.
For example, one of the data marts is a daily census cube that allows the hospital to see over the course of a year on a daily basis what the occupancy is like at each of its five units. The cube gives managers the ability to identify trends in bed allocation and plan staffing loads accordingly.
Another impressive example is a conservable length-of-stay analysis, a huge study that Emmons says would have cost a fortune using traditional methods. Using the QUIST system to create the report, hospital management was able to compare all of its actual lengths of stay against case mix matched lengths of stay and identify potential areas for improvement across the board at the level of the individual health unit. The study will help management identify conservable days and, where possible, reduce lengths of stay to better accommodate more patients within the hospital using the same resources.
Emmons says that approximately three-quarters of her team’s task is now complete. A user-satisfaction questionnaire has been developed to determine whether users know about the system, have used it, and like it. The plan is to continue constructing more cubes and refining existing work until Christmas when QUIST will look at bringing the other hospitals comprising VHHSC onto the system. According to the strategic plan the team is then scheduled to turn its attention for the next two years toward enhancing decision support.
“Now that we’ve pulled these wonderful, rich data-sets together our intent is to go out and teach people how to use the data and how to ask the right questions. We want to facilitate decisions based on information. The more information you create the more wisdom you facilitate,” stresses Emmons.
A FINAL ASSESSMENT
Emmons believes that the tactical approach is a good idea, even if an organization plans on creating a data warehouse in the long run. One of the main reasons: building data cubes helps one identify the information that’s going to be really useful.
“I don’t truly believe we would have really achieved anything if we had taken a longer-term strategic approach,” she surmises. “By the time we would have had something to deliver our users would have been too frustrated to use it. Even worse, management probably would have pulled the plug on the system because unless it provides valuable information quickly, the development costs are nothing but overhead.”
Thankfully, QUIST is in the enviable position of never having to second-guess its decision. It’s simply too busy helping hospital management ask the right questions, and in so doing, improving the quality of health care.
Pat Atkinson is a Toronto-based freelance writer and editor.