Like many integrated delivery networks (IDNs), Hunterdon Healthcare System has worked over the years with a varied group of software products developed to provide the business intelligence and analytics needed for optimal decision making. The inevitable result has been generation of highly user-friendly dashboard graphics that are exceedingly attractive but of very limited practical use.
What happens between application concept and delivery into the healthcare environment? Without any exception, these decision support-system (DSS) products require storing information on a separate database, external to the application, and then running batch jobs every night that copy the data to the dashboard database for viewing. As a result, the information that appears on the dashboards is often out of date, fragmented and incomplete.
Despite these less-than-optimal results, multiple driving forces combined to ensure that the search continue for the right DSS technology – one that enabled rapid development and delivered information in real-time mode. Earlier this year, Hunterdon began the rollout of a new hospital performance-management (HPM) system that utilizes InterSystems’ DeepSee embedded real-time analytics technology. Based on early results and end-user feedback, this system succeeds where previous DSS approaches failed in delivering the in-depth, easily usable information Hunterdon needs to address expanding reporting requirements and support optimal care delivery.
Driving the search for real-time analytics
Based in Flemington, N.J., the flagship of the Hunterdon IDN is the Hunterdon Medical Center, a 178-bed teaching hospital that provides a full range of preventive, diagnostic and therapeutic health services. Hunterdon treats more than 8,600 in-patients annually, with 33,000 emergency department visits and more than 292,000 outpatient visits per year.
Typical of many hospital facilities, Hunterdon has taken a best-of-breed approach when selecting and implementing healthcare technology solutions. As a result, the IT department supports a variety of applications with hundreds of interfaces from multiple vendors, making it a major challenge to normalize data across systems and view information in a timely way. In this scenario, viewing information at a strategic level required a manual process where staff members were charged with pulling data from multiple applications and building spreadsheets that combined relevant data to provide a high-level view. The growing demand for information at a strategic level was a major driver behind our ongoing search for a new DSS application. And that demand is coming from personnel in all organizational areas, including clinicians, administrators, financial specialists and C-suite executives.
Adding to the challenge is that reporting at local, state and federal levels is becoming more complex than ever. Hunterdon, for example, receives report requests from 29 external agencies. The large number of information requests streaming in from government entities tests our reporting capabilities from the perspective of sheer volume. Just as critical is the fact that data definition is by no means consistent in the area of regulatory reporting. For example, one measure often applied to surgical processes is the amount of time required for a particular procedure. While one agency may be asking for the time elapsed from when the patient enters the operating room until he or she is wheeled out, another may be asking for when surgery actually began and when it ended. These inconsistencies forced data collection into separate silos and had a negative impact on overall data integrity.
All of these factors contributed to Hunterdon’s decision to implement the HPM system.
Overcoming the hurdles
Not too surprisingly, the main roadblocks to successful implementation of real-time analytics – as is true of many enterprise-scale systems – come from people rather than from technology. At Hunterdon, for example, it was difficult to realize and accept that data integrity was not at the level we had previously assumed. Neither technologists nor end users realize exactly how much data massaging and adjustment is taking place when information from separate systems combines to deliver a strategic view until they see the information in real time.
There was some pushback from those who had tried to use other dashboards from previous implementations and found the information delivered to be less than useful.
These challenges are gradually disappearing as executives realize and accept that the data itself, rather than the system, is a major issue, and that real-time analytics makes it possible to understand and effectively utilize that data to drive informed decisions and improve outcomes. Equally important, we are discovering that embedded DeepSee analytics can help uncover missed opportunities that are inevitable with purely retrospective analysis – the only analysis that is possible when data is being stored in a separate data warehouse and gathered for reporting via batch jobs run nightly rather than in real time.
Early-stage benefits delivery
Hunterdon began the HPM rollout in the first quarter of 2012, which means that any attempt at benefits quantification would be premature. At present, the focus is on measuring certain specific indicators, including:
- Readmission indicators for pneumonia, heart failure and acute myocardial infarction;
- Clinical indicators for patients with obstetric trauma, vaginal delivery with and without instruments, patients with postoperative respiratory failure and mortality during a hospital stay; and
- Lengths of a patient’s hospital stay by physician, by certain diagnoses and by payer.
Decisions concerning what should be measured and how analytics should be displayed are made by a team that includes executive/financial management, clinicians and IT specialists. It’s already evident that the information being delivered is going to prove very useful for responding to the ever-expanding reporting requirements of government agencies. After stopping reimbursement for preventable medical errors, for example, the Center for Medicare and Medicaid Services (CMS) is now focusing on preventable readmissions and measuring readmissions in the Medicare fee-for-service program. In 2011, the National Committee for Quality Assurance added an all-cause readmission category to its measures, and in the second quarter of 2012, the National Quality Forum endorsed those metrics. While there is a fair amount of ongoing debate about the relationship between hospital readmissions and quality of care delivery, Hunterdon is now in a better position to respond to agency information requests with accurate data delivered in a timely manner.
Although evidence is anecdotal at this early stage of analytics implementation, there is agreement among Hunterdon clinicians that the ability to access patient information on a continuous basis will enable improved care delivery and outcomes.
Recommendations for laying out the analytics roadmap
Selecting the appropriate software product for analytics-based decision support is a critical success factor. However, before that decision is made, it is essential to ensure that the right people in the healthcare organization are involved from the start of the initiative. It is strongly recommended that the CIO identify the executives, clinicians and staff members who will benefit most from a real-time analytics dashboard system. Once there is a clear understanding of the pain points of potential end users, it becomes easier to make them your champions within the organization.
When there is sufficient internal support for the initiative, the importance of selecting the right technology to provide the information foundation for the dashboards cannot be overstated. DeepSee proved to enable fast development, supported very rapid information retrieval and eliminated any need for a separate data repository. In addition, this technology choice made it possible to leverage Hunterdon’s existing investment in the InterSystems database system and integration/development platform. InterSystems provided strategic support throughout the HPM project and recommended Cognizant, the consulting firm that teamed with Hunterdon personnel to develop and implement real-time analytics in our environment.
In summary, building internal support throughout the organization, closely examining the technology foundation of any DSS solution and pulling together an experienced development team are keys to successful analytics implementation. It means a significant investment of multiple resources, but Hunterdon is convinced that the payoff in terms of improved decision making and optimal care delivery makes that investment worthwhile.
About the author
Glenn Mamary is vice president and CIO at Hunterdon Healthcare. For more on InterSystems, go to http://www.intersystems.com/industry/healthcare/index.html.
