Healthcare Analytics Applications: Why You Need an Out of Box Solution with Customizability
Health systems throughout the U.S. are facing tough decisions about which healthcare analytics applications will be best for them. They know they need to unlock the value of their data to help ease their reporting burden for government mandates, to provide better care, and to decrease the cost of care.
Decisions become even more difficult because of new electronic health record (EHR) implementations, other IT projects, and the complexity of the existing IT data infrastructure. Add the constantly changing nature of the healthcare environment — whether regulatory changes or advances in medicine and technology — and the task to weigh the pros and cons between the many healthcare analytics offerings on the marketplace becomes quite daunting.
In order to meet these demands, analytics solutions must be able to provide quick time to value in months rather than years. They must also be customizable to provide the in-depth analytics needs necessary to support today’s ever-evolving clinical world. With a late-binding architecture, such a solution is possible.
Purchasing an analytics solution: customizable or out-of-the-box?
Purchasing an analytics solution can be compared to purchasing a model home: every model home is part of a standard package and will provide an acceptable, functional living space for the majority of home purchasers. If the home purchaser has specific needs and wants, however, they can add customization and upgrades to the basic home model.
Many of today’s analytics solutions offer standard features, just like the standard package of a model home, which are applicable to most health systems’ analytics needs. There’s a major benefit to these off-the-shelf offerings: the time to value. With a fast implementation time and the ability to provide quick answers to basic problems, waiting for years before realizing a return on investment (ROI) is no longer necessary. Instead, an ROI can be realized within weeks or months.
But what if the basic home model doesn’t provide the home owner with the best design they need for their living situation? Being able to easily add custom features to the home with the original builder is similar to opting for an analytics application that can be easily customized, something made possible with a late-binding architecture.
Why customization is essential in today’s clinical world
To really gain a deep understanding of the organization and its patients, customization of the analytics application is necessary. Customization enables analysts to dig much deeper into the data — and not just after the initial implementation. Instead, the best type of customized healthcare analytics applications can accommodate endless customizations time after time based on new definitions and rules. This design enables the healthcare organization to evolve by asking questions and providing answers that are actionable and applicable to the specific organization which ultimate leads to new questions as the answers from previous questions provide new insights.
An example of this need to customize an analytics platform can be found at one of our clients that is a children’s hospital. Many commercial analytics applications treat children like short adults. But children have many additional criteria, such as growth progress and immunization history, that must be tracked. Even the units of measure can be different. A customized analytics solution can be built to take those differences into consideration, so the results are appropriate and relevant to the patient population being served. To complicate the need to customize an application to be child centric, throw in the unique regional requirement of the hospital to support the clinical treatment of poisonous snake bites experienced by these children. Without the ability to customize these analytics, very few “out-of-the-box” solutions would be able to meet the requirements of this special and unique patient population.
Customization, however, isn’t just about adjusting to different patient needs. Each health system has its own workflows and data requirements. Even two hospitals using the same EHR will likely have differences in the way they implemented it. With a customized solution, it’s possible to mold the analytics to the best clinical processes and workflows defined by a particular clinical group. As those workflows are implemented, the customized analytics can measure the effectiveness of the defined workflows. If the analytics uncover a better workflow, workflows can be altered, and the analytics once again can be customized to measure the new workflows creating a closed feedback loop. For example, clinical staff can now receive a standard order set for sepsis regardless of which department they’re in. If, however, a new standard workflow is clinically proven, the analytics can be customized once again to measure the efficacy as well as monitor the variability of the workflow within the hospital.
Lastly, as hospitals and health systems join accountable care organizations (ACOs) and/or take on more risk, they’ll want to be able to customize how they look at patients under the ACO model versus those still under the fee-for-service model. Patients under the ACO model need to be managed differently, with greater focus on population health management. As a result, many new parameters previously ignored or at least of minimal interest now need to be taken into consideration. Customization allows for these new considerations — and to make adjustments more easily as the requirements and realities of the healthcare industry continue to change.
Key to customization: Late-Binding™ architecture
One of the keys to enabling this level of customization and flexibility is the ability to work with data at the atomic level without transforming it, which is the unique characteristic of a late-binding enterprise data warehouse (EDW).
Typically analytics applications rely on early-binding methodologies and defining business rules early in the data storage process, which limits their ability to be altered to accommodate new information without extensive time and labor. In contrast, an analytics application solution designed with a late-binding architecture, leaves data unbound in the source marts in its original format until it’s needed by the customized analytics application. As a result, colleagues in different departments, such as clinical, research, or finance, can take the same core data from the EDW, apply different rules to it, and use it to perform analyses that are vastly different. Furthermore, as regulations, rules, or best practices change, metrics can be changed or created since the source mart data is still in its atomic form.
This doesn’t mean everyone in the organization has free reign to do whatever they want. There still has to be a level of governance at the top that oversees standardization of how data is managed at the enterprise level while also approving exceptions. That is especially true when certain departments are required to deliver data according to specific requirements, such as those from the Centers for Medicare and Medicaid Services (CMS).
An example we’ve seen of this type of flexibility is one department calculating patient days differently from the rest of the hospital. If there’s a good reason, then this is fine, but someone has to make the determination on whether the reason is good enough. Proper data governance will ensure the system can accommodate both standardization and selected deviation.
Getting it right
With the growing importance of analytics in the healthcare industry, particularly in light of population health management concerns driven by the shift to pay for performance, speed is important. But it’s not the only factor. Speed related to short-term needs must be balanced against the demand for more organization-specific information and longer-term solutions.
The ability to create customized healthcare analytics applications, and enhance or adjust them easily as needs change, is critical to using the valuable information in the clinical, financial, and other source applications to make better decisions. Since healthcare continues to evolve, advance, and transform, it only makes sense to choose a healthcare analytics application with the core analytics requirements as well as the ability to accommodate guaranteed changes.
If you’re using out-of-the-box solution for your analytics needs, do you foresee a time when you will need to customize your analytics applications? Do your out-of-the-box analytics applications allow for customization? How would your health system use a customizable analytics application to ignite actionable change?