Let Them Eat KPI

The spotlight on IT in the healthcare industry has never been more blinding. Healthcare reform (e.g., Meaningful Use, ICD-10 conversion) has mandated IT investment in the name of improved clinical/quality outcomes, reduced medical errors and standardized care. Even without these reforms, however, the increasing role of IT in positively impacting the cost and quality of patient care has become a top priority for healthcare provider organizations.

KPI is King

At the center of demonstrating – and measuring – success is the Key Performance Indicator (KPI).  The healthcare industry has a very large number of complex KPI.  From research to patient care to operations and finance, KPIs and their management (metadata management, master data management) are critical to overall organizational success.

But with increased reliance on these comes increased complexity and costs, both in creation and management of all this data.*

(*Do not confuse “all this data” with “big data.”  As told in the story of an online bookseller’s success, a big data approach involves storing everything first and interpreting later. The KPI model, in traditional data management style, identifies desired data first and stores that data in context.)

Creating a KPI Monster

Because of the sheer volume of KPIs in use, as well as the complexity of calculating even one KPI, the intricacy of the KPI world should not be underestimated.

Consider the many factors to be addressed in determining “ER waiting time”: information about each patient, symptoms they present with, severity of the condition, real-time calculations about number of patients in various stages of treatment, and the list goes on.

Herein lies the potential for creating a monster.

A KPI is very far downstream in the business objectives pipeline, and the work involved in calculating any KPI runs the gamut from validating the integrity of the data, to executing the data product the KPI and its underlying data live in, to managing systems for analyzing and reporting on that KPI. This lifecycle includes:

  • Managing the sources of the data
  • Data quality and cleansing
  • Managing the definition of the data elements used in the KPI
  • Tracking ownership of the KPI and data element definition
  • Metadata management
  • Transparent definition of the calculations used to create the KPI
  • Maintaining KPI “trendability”

Managing KPIs, particularly in healthcare, is extremely complex. The healthcare industry has more stakeholders for its KPIs than is typical in most corporations, and each of those groups may be interested in different metadata about those KPIs. For instance, “over 75 years of age” and “super senior” might be compatible in spoken language, but have a completely different taxonomy when expressed in metadata.

Often, however, the IT systems responsible for KPI management are already overly complex and serve too many disparate functions, as they’ve usually evolved over a long period of time by many, many people.

In turn, the IT department’s management of these systems is hyper-focused on routine execution of processes that could and should be automated, and ends up expending its resources on troubleshooting and supporting. Statisticians, who ideally evaluate the trends and spot opportunities for improvement, may spend more time manually calculating standard statistics than doing the value-added analysis they were hired to do.

The prospect of adding new KPIs means the exponential addition of – potentially – thousands of data points, and incorporation into an IT system that is already overburdened.  And with an exponential increase in complexity comes an exponential increase in management costs.

Where KPI is King and MDM is Hero

The underlying goal in a KPI model is to create a structured, common language, understandable to all of your stakeholders, to communicate what’s supposed to happen and then build execution engines to translate that into actual activity. Welcome to Master Data Management (MDM).

Creating a common language means developing an XML (Extensible Markup Language) – whose words and letters are always pronounced (defined) exactly the same – and there is no guesswork, unlike, say the English language, which is full of pronunciation exceptions. So “super senior” or “redhead” would ALWAYS mean exactly the same thing to all stakeholders.

Now that all parties, even IT, understand and speak the same language, life for all stakeholders becomes less about tedious maintenance and troubleshooting, and more about value-add and better outcomes:

  1. The potential for interpretation errors between the groups is dramatically minimized.
  2. All stakeholders can have their ownership and interest enforced on the KPI.
  3. KPI is “trendable” and reliable over time.
  4. The underlying technology that executes the KPI-MDM pipeline can be updated with little or no need for bringing together the disparate parties to redevelop the definition of the KPI.
  5. As much as possible, statisticians and researchers are empowered to specify and own their KPI.

The beauty of a properly built and executed KPI strategy is that you can leverage the talent and skills you already have using well-proven technology. There’s no compelling reason to invest in risky, cutting edge technology because you already know exactly what you need. (Again – highlighting the separation of this discussion from “big data.”)

Are you ready to create a KPI-MDM strategy that can be ported into the future?  This readiness starts with the understanding that an investment into IT infrastructure has proven ROI. The KPI model underscores the value IT brings to healthcare – and any industry whose outcomes can be objectively measured.

Between government mandated healthcare improvements and internal, mission-based commitments to providing quality health care, healthcare providers are recognizing both the need, and the opportunity to deliver better results through IT.

 

Image Source

Comments are closed.