- About Us
- Centers of Excellence
- Case Studies
- White Papers
- Mobile Solutions for Improving the Patient Experience – Managing Wait Times
- 6 Considerations for Successfully Managing a Meaningful Use Program
- Start Driving Value from your EHR with Eight Real World Lessons Learned
- Make It Happen
- Optimizing the Ambulatory EHR – A Systematic Approach to Realizing Value
- Physician Leaders’ Role in IT Governance
- Building a Technology Roadmap that Supports YOUR Organization’s Value-Based Model
- Meeting Meaningful Use Patient Access Requirements - A Mobile Solution?
- Evolving Role of the CIO = The Synergist
- No Easy Wins with an EHR
- Are We There Yet? Realizing Value from an EHR
- Be The One - A Single Accountable Leader
- A New Path to Physician Engagement
- One Approach to Population Health Management: Start in Your Own Backyard
- Aspen's work with Baystate Health to realiz value through an ambulatory EMR optimization
- Extending Your EHR to Independent Hospitals and Clinics: Five Considerations
- Solving the Specialty EMR Implementation Puzzle
- Provider Considerations for ICD-10 Delay: Don’t Take Your Foot Off the Gas
- Round Peg in a Square Hole: Behavioral Health and EMRs
- EMR Adoption Best Practices
- How to Fast-Track Your Meaningful Use Effort
- The CIO-CMIO Partnership: A Recipe for Success
- EMR Dress Rehearsal Best Practices
- IT Governance Best Practices
- 10 Lessons Learned for Achieving the Patient Safety Improvements of Biomedical Device Integration
- Meaningful Use Security Risk Analysis
- User Provisioning Best Practices
- How to Achieve an ROI for Healthcare BI
- Three Steps IT Should Take in Response to Healthcare Reform
- iPad: A Tablet Computing Renaissance in Healthcare?
- A Field Guide to Online Personal Health Records
- Optimizing the Business and IT Relationship
- Using SSO to Streamline Security and the Clinical Experience
- Cloud Computing in Healthcare: Is there a Silver Lining?
- Great Project Management = IT Success
- Guidelines for IT Planning for Healthcare Construction Projects
How to Achieve an ROI for Healthcare BI
By Jim B-Reay, Aspen Advisors
As hospitals make significant investments in electronic medical record (EMR) technology, along with related updates to hospital billing, materials management, costing, and quality systems, they typically find that the promised analytics and reporting are not adequate. To tie together data from these disparate systems and even to optimize access to data within an integrated system, a Business Intelligence (BI) strategy is needed.
A typical BI strategy encompasses data governance; data staging and warehousing; tools for query, reporting, and dashboards; and a staffing model to build the initial framework and expand the architecture to serve the changing needs of the business.
For many organizations, this additional investment is hard to justify considering the outlays already made in their core systems. While working on one strategy recently, I was asked: “If we make this investment, how can we measure the direct return on investment (ROI)? What is the actual ROI of an investment in BI?”
To help the client answer these questions, I reached out to a dozen organizations, all of which have BI programs of some degree of maturity and asked the very same questions. The responses I got were different and enlightening. I found that successful sites had a common theme: BI value is based on the use of the system to analyze data from various clinical and administrative systems and the willingness of the organization to act upon the findings to make changes that ultimately improve productivity and efficiency.
The twelve participating organizations included medium-sized facilities with 200 to400 beds and mid-size and larger IDNs with six to 12 hospitals. All have EMRs from commercial vendors including Epic, Cerner, and Meditech. They vary in the degree to which their ambulatory data is integrated.
Technologies and Tools
Their BI technologies vary too. Some have mature home-grown data warehouses; several are using older integrated reporting systems like TSi and Trendstar (and their newer versions, EPSi and McKesson HPM/HBI). Other sites are implementing new data warehousing appliance products. Many are relying heavily on their EMR-provided reporting databases, and others are bypassing the data warehouse entirely and using in-memory aggregation tools like Qlikview and Dimensional Insight. For presentation, many use reporting tools from Cognos or SAP/BO, and others have custom developed web-based dashboards.
While these organizations vary in size, EMR maturity, and technology, I found commonalities in their responses.
A Cost of Doing Business
Many of the respondents stated that there wasn’t a planned ROI; they saw the investment in BI as a cost of doing business and considered BI as a necessary investment for which the value would be proven using the results from the analytics. Thus, they did not establish clear financial goals beforehand. Instead, they identified gaps in their data environment that a BI strategy would address and chartered projects to suit. Several respondents said “we are only starting our initiative, so we won’t know about our results for a while – check back next year!”
The organizations with the cost-of-doing-business viewpoint are pursuing a number of strategies:
- One is building a major data warehouse to aggregate data from multiple EMRs.
- Two are implementing “shrink wrap” data warehouse solutions developed at major academic institutions.
- Several are implementing “offsite data aggregator” solutions in which dashboards are delivered based on claims and accounts receivable feeds. However, the data latency in this particular strategy, up to 45 days in one case, may render the results of little use to take corrective action in a timely manner.
- Yet another is “rebooting” a stale data warehouse that had failed to deliver tangible results in the past.
While there is no question that having a solid BI foundation will unlock potential value, the sites that are taking the cost-of-doing business approach – also known as the “build it and they will come” approach – need to make sure they start with the issues to be addressed and then build the governance and associated projects. Having worked in BI since the mid- 1990s, I have seen more than a few ambitious projects fail because they weren’t built to address specific needs but were purely architectural, technology marvels.
The sites that are implementing packaged solutions will need to make sure that the delivered metrics are relevant to their needs and that the trade-offs inherent in some tools are acceptable for their organizations.
Empower the Analysts (Plus a Little Insurance)
A smaller group of the participating organizations had a slightly clearer idea of what they’re trying to achieve with their BI investment: empowering their data and business analysts. In these cases, the organizations have fairly seasoned analysts who are clamoring for better tools to continue their roles as data analyzers.
In most cases, these analysts had established their value years earlier, able to conjure data from a variety of sources and combine it into water-cooled spreadsheets or custom Access databases. The stated goal of the BI investment was to give these analysts more robust tools with which to enhance the great work they were already doing and the value they were already providing to the organization.
This approach drives to more standardization of data and allows for replication of the current mysterious data manipulations of these trusted analysts. In addition, replacing the desktop database with an IT-maintained warehouse and a heavily macro-filled spreadsheet with a set of summary tables and dashboards provides a measure of insurance that the knowledge and analytics would be securely in place should the analyst decide to move on or could be used by others within the organization.
There are positive aspects to this approach. The organizations have a clear vision of what benefits they will be getting from the investment and realize the value in standardizing poorly documented, one-person processes. The challenge in finding ROI is that these analysts tend to work very hard for very reasonable compensation, usually with deep tenure, and may not be leaving anytime soon. A cynical view may be that by taking this approach, you’re investing dollars in something that the analyst was already providing and likely would provide for years to come. However, these organizations are enhancing the capabilities of their trusted analysts and expanding what they can provide in addition to “buying insurance” for the future.
Targeted and Tactical
A core group of respondents challenged the premise of BI ROI by saying that BI has NO value to the organization in and of itself unless the project is matched to strategic initiatives. Their BI projects, interestingly enough, were often much smaller than the “insurance” or “build it and they will come” initiatives.
One hospital, for example, built a Microsoft Access-based datamart that received data on a weekly basis from six different systems. The feeds were all small with a limited set of data and all targeted to achieve a single goal of analyzing and optimizing physician outpatient schedules. The results were used weekly to rebalance schedules and find slots to shorten wait times for appointments and monthly by leadership to assess strategic hiring to staff up on higher volume areas.
Other facilities were able to identify similar limited scope datamarts that offered direct ROI including a group that did targeted analysis on disease state outcomes and another that used BI tools to build near real-time core measures analysis dashboards.
In all of these cases, there was a level of BI infrastructure required to make this all work, but the level of direct investment required was, in most cases, far less than a full soup-to-nuts data warehousing initiative. The ROI realized was the result of targeted, limited scope initiatives with only just enough infrastructure to deliver these results.
It is worth noting that one of these respondents is a large academic medical center with a mature data warehouse and large data warehousing team, but the targeted projects in many cases actually bypassed the core data warehouse team and database. At a major pediatric institution, they actually canceled a major re-vamp of their core warehouse, restructuring their team toward targeted delivery teams with specific ROI-driven charters.
The challenge going forward with the targeted and tactical “ROI per project” model is scalability and standardization across the organization. It is important to have a small data management team who pays attention to the overall data model and analyst teams that consistently use the shared data model to reduce data redundancy.
Although there were a few cases where it appeared that investments were being made to get BI in the door without truly understanding the solution on offer, those that had embarked on their BI strategies with a solid set of requirements and strong governance will be well served by their investment. There are complex questions that these organizations simply would not be able to answer without the data aggregation and query toolsets that an investment in BI brings.
But direct calculation of a return on investment can be difficult. For the “build it and they will come” group, they have made it clear they’re willing to let the ROI be determined through later projects. What the third group of respondents showed was that if you’re looking for ROI, you need a clear definition of scope and the organizational ability to respond to findings. It is possible to get an amazing ROI from a project with one smart analyst, some extract files, and an Access database. But it’s up to the organization to take that information and act on it, and it’s up to IT to build a support structure to ensure that that information continues to be available.
What My Client Learned
In light of these findings, we studied my client’s five year investment plan for BI initiatives and found that of the three types, they were setting off in more of a “build it and they will come” model. Since this model is the hardest to define an ROI for, we restructured the plan to better match the third category – delivering strong ROI through targeted initiatives but building towards a solid BI foundation.
Working with leadership, I encouraged them to find some smaller, targeted initiatives – projects, for example, from the CMO and CNO that had limited scope but measurable returns identified. We mapped those into their planned investment. This changed the five year plan to have a limited budget for software and infrastructure in the first two years. There was no point in purchasing enterprise toolsets at that stage. The funding was earmarked instead for desktop analytics tools to support the budgeted analysts, who would be targeted toward the smaller projects. We scheduled the larger investment in years three, four, and five to help “productionize” these smaller projects and slot them into a larger architecture.
Best Practice for Delivering ROI
To design and implement a Business Intelligence initiative that delivers a positive ROI, start out with a limited scope and strong organizational support for acting on the findings. Select a single study area, get clinical support, and assign the most experienced analysts (second model) with support for data extracts as needed. Once you have proven value to the organization, look for ways to expand. Work to productionize the extracts and move the database off of the analyst’s desktop, so the value you get from that first study area is preserved and re-useable. Work on back-loading additional data as needed to expand the study area.
Find a second and third related organizational problem that could be piggybacked on the dataset you’re using and find an organizational sponsor who will take the action needed based on the BI data findings. If possible, expand the existing structures to contain the data needed for the new studies, but don’t create a tortured data model. Don’t be afraid to create another targeted data mart as needed.
In parallel with this first initiative, start building strong BI governance in the organization. Ensure that analysts across the organization are meeting regularly to discuss and document data standards and that wheel-reinvention is minimized. This can be a matrixed group rather than a formal reporting organization, but participation needs to be mandatory. The lead for this analyst group should be invited to executive-level steering meetings to listen for areas of frustration and concern with data and be able to both represent the work that is being done and bring the concerns back to the analyst team for action.
Through targeted initiatives, experienced analysts, and strong governance, BI projects will have a tangible ROI.