Thanks to web 2.0 and ubiquitous social websites, the amount of information available on the Internet is exploding. The same can be said for enterprise applications-data being gathered and generated across business applications are exploding, too. When looked at through different lenses, this untapped data, often thought a liability, can manifest itself into information and/or intelligence for an organization.
The key is managing and understanding this data in an organizational context and turning it from a perceived liability into an asset. This holds true for almost all industry verticals, including healthcare. An analogy might be to consider data in an application to that of blood coursing through a body. An application can live and thrive as long as data flows in and out of it.
Like blood, data are irreplaceable assets. It's also time sensitive and leaves a trace behind. If you look closely, data can be used as a strategic asset that can be leveraged to create long-range business strategies and introduce efficiencies to an organization. Most organizations know the value of data but have a hard time managing and governing it; or in some cases, the value of data management is not realized until initiatives like analytics or enterprise data warehousing are taken seriously.
It's important to note that more then half of analytics projects fail due to poor data quality. Poor enterprise level data management can be visible at organizational level reference data, such as departments and employees; that is, financial, clinical, and revenue systems each have their own names and identifiers for departments and employees.
If this sounds familiar, rest assured that you are not the only one with this problem; it's no wonder millions of dollars are invested in dashboards and reporting systems that don't work. Management is continually looking for that killer ‘report’ that can give a 360 view of an organization, but unfortunately, data management and quality problems can't be solved by reporting systems.
Perhaps the answer lies in managing data at the enterprise level. Understanding enterprise data could be the key that unlocks the door to a detailed view of an organization's systems. It provides a strategic, horizontal view of information across the enterprise. It helps identify common threads, objects, and information being used in cross-departmental functions.
For example, lab results originating from lab systems flow into the health information system (HIS), and along the way, doctor's notes get incorporated. This data could then be used by physicians for decision-making, reporting to healthcare agencies, infection control, quality improvement, and, sometimes, clinical research. This simple set of data contains a significant amount of information and intelligence, and when analyzed over an extended period of time, it could prove to be an indispensable source of clinical analytics.
Where are enterprise data? Data life begins when applications that are built or acquired go into the production environment. Most healthcare providers depend on off-the-shelf products or applications these days; hence their focus is primarily application integration.
For example, vendors may provide a list of HL7 feeds they want to receive from an HIS, and once those feeds are delivered, the application is implemented, and life is good. The project management office is satisfied because one more project can be checked off its list.
But keep in mind that on the backs of these applications runs the data engine that keeps humming along churning out data. Logs and archives are a great source of hidden information and data. These computer systems or business applications run a variety of database managements systems (DBMs); they could be file-based DBMs, spreadsheets, or relational table-based systems such as Oracle, SQL Server, or MySQL. If you happen to be a typical hospital with 100 or so applications, your data are living in those hundreds of different DBMs or in spreadsheets or in some database administrator's brain.
Your database administrator might be able to tell you where the data are depending on how the data are managed across the organization, but he might not know the context and purpose of those data elements. If you happen to be one of the 5 percent of healthcare providers with enterprise level metadata and a data steward, you are in an advantageous position; but for the other 95 percent with meaning, purpose, and context of data hidden in the brains of long-term employees, you are in a disadvantaged position.
BUILDING A REGISTRY
For those wanting to take an organization to the next level of data maturity, having an enterprise level meta-data registry is a key step. Enterprise level data management is a monolithic project and task, and while IT organizations usually do a good job in maintaining and understating their networks and servers, the same can't be said for data.
Enterprise level data models are uncommon in the healthcare industry, and it can take a long time to build one from scratch. For an organization looking to build an enterprise level meta-data and data model, a practical strategy may be to start with a high ROI data hot spot or a high visibility data specific project. If your organization is looking into revenue management and optimization, this is a wonderful opportunity to start data modeling and management from these subject areas. At the end of the exercise, a clear view of the organization and its data flow should emerge.
When should an organization start building an enterprise level meta-data registry? The answer is now. With meaningful use and the emphasis on health information exchange, data will become increasingly important to an organization. Poor data quality will get noticed and new waves of data projects will begin to surface, but why be reactionary when you can be proactive and start getting results in the near future.
Accurate data can ease the burden on business units, IT, and management. Moreover, data sharing works only if data quality and integrity is maintained, and data quality and integrity work only if a solid enterprise data strategy and framework exists. Integrating data governance and data management as part of project management methodology is a conceivable way to entice project managers to consider, utilize, and maintain the integrity of data.
Remember, just like the blood coursing through our veins, data generation never stops. It's time proactive organizations take a look at building an enterprise level meta-data registry to harness this “unsung” asset-data.
Ajay Khandelwal is manager of enterprise architecture, Children's Hospital Los Angeles. Healthcare Informatics 2011 August;28(8):52-54