In the third chapter of our special report on healthcare business intelligence (BI), consultant Lou Agosta provides...
BI project management tips for healthcare organizations, including advice on assembling a BI team and evaluating BI software.
Table of Contents
Healthcare business intelligence systems: an IT laggard no more?
Healthcare BI software in action: real-world examples, best practices
Business intelligence project management tips for a healthcare BI team
Key IT roles on healthcare business intelligence project teams
Business intelligence (BI) projects are distinct among IT initiatives in that they address questions that look beyond the day-to-day operations of an organization. BI systems are designed to enable users to track and analyze trends on internal costs, revenues, resource utilization and market dynamics – and in the specific case of healthcare BI applications, they can help providers measure and assess a variety of treatment and healthcare quality metrics in an effort to improve patient care.
From a “work profile” perspective, BI systems tend to be query-intensive instead of update-intensive, as transactional systems are. Large volumes of data generated over time get aggregated in a data warehouse or data marts; reports are produced for some users, while others (business analysts, etc.) are given BI tools that let them execute ad hoc queries against the accumulated data. But first, of course, the data warehouse has to be built and the BI software needs to be deployed. And the distinct nature of BI systems poses unique implementation challenges for BI project managers and their teams.
In healthcare as well as other markets, BI projects are exposed to a whole set of potential “gotchas” and mistakes. Don’t assume, for example, that the data being provided by transactional systems is accurate or consistent from system to system. Investigation of operational systems often shows that the data stored in them has never been entirely accurate and may even have been systematically skewed. This can raise significant inter-department issues and conflicts that BI project managers need to referee or call to the attention of their executive sponsors for higher-level rulings. In addition, substantial data cleansing efforts likely will be necessary. Making sure that you end up with accurate data is especially important for business intelligence in healthcare, since your BI system likely will be used to support clinical decision-making, patient-care analysis and other truly mission-critical applications.
Getting started on business intelligence project management
A good place to start, then, is with a readiness assessment specifically targeting the data sources that will be used to populate the BI system. It’s also wise to conduct design reviews early and often and get input from end users on the critical issues they face in their jobs and the specific questions they want the BI software to answer for them, as well as the usability of the client interface. These assessments and reviews should help you bring major shortcomings and omissions to the surface and avoid big surprises later in the project. Issues that are addressed in the requirements and design phase are an order of magnitude less expensive to correct than those that come to light at implementation time.
Political issues also can loom large on BI projects because of the data integration requirements. Some departments may be territorial about “their” data and reluctant to share it with others. While the data security requirements imposed by laws such as the federal Health Insurance Portability and Accountability Act (HIPAA) are necessarily a priority on healthcare BI deployments, don’t let them become an excuse for lack of cooperation within your organization. As long as the data in BI systems is kept inside corporate walls and used to provide and/or improve care for patients (who, after all, have asked for help), HIPAA shouldn’t be an obstacle – either a real or a politically inspired one.
Assembling a cross-functional team to engage these challenges from the perspective of the entire enterprise is a BI project management best practice. As mentioned above, the BI team and the BI project as a whole typically require the sponsorship of the CEO, CIO or another powerful executive who can ensure that funding is available and to whom disputes can be escalated if necessary.
BI project management and the business intelligence competency center
Not every organization will want to centralize its BI resources and set up a business intelligence competency center (BICC) or a BI center of excellence. However, for those that are so inclined, a cross-functional team set up for an initial project can become the nucleus of a BICC that would be responsible for BI project management and for proliferating BI best practices throughout the organization.
Managing vendor relationships is a full-time endeavor – and not one that can be restricted to the IT procurement department, especially if the software is mission-critical, as healthcare BI tools tend to be. A wise proverb holds that “it takes a whole village to raise a child.” Well, it requires a whole enterprise to get value from BI software. Early involvement of the end users who will actually be “hands on” with the BI system is essential to the ultimate success of a BI project.
During the software evaluation phase, go to BI vendors with a written checklist in hand and ask about the software features and functionality that are important to you and your end users. You also should check references at other organizations that have done healthcare BI business with the vendors you’re considering. Live demos at a client reference site, vendor lab or both require time and effort but usually are worth it in the long run. Negotiating a written service-level agreement (SLA) is another key step.
Finally, while the recession may officially be over, it’s still a buyer’s market for IT, including BI tools. Vendors are hungry for revenue and discounting is at the point of no return. Still, wise end-user organizations understand that vendors have to hit their numbers, too, and are willing to negotiate for extras such as free training and enhanced SLAs, not just straight purchase-price discounts.
About the author: Lou Agosta, Ph.D., is an independent industry analyst specializing in data warehousing, data quality, data mining, business intelligence and healthcare IT. Keyword: data. Agosta's book, The Essential Guide to Data Warehousing, was published by Prentice Hall PTR. He can be reached at LAgosta@acm.org.