Your Enterprise Data Warehouse: Why it will take a Village to Build an EDW

http://www.dreamstime.com/royalty-free-stock-photos-village-manarola-ferry-cinque-terre-italy-europe-image45312618

In the previous blog posts we’ve talked about how yes, you can do more with less by employing the proper tools for the job. But that will only get you so far. The requests for new reports, analysis, and even whole new datasets are coming too fast and furious to keep up. Gone are the days when bringing billing and administrative data into an EDW and throwing report and ad-hoc query tools on top was enough. Now we need cardiac registries and patient satisfaction and centralized metric repositories brought into the mix. So no, we can’t do all of this with less, but we might be able to do it without growing the EDW/BI team or at least without growing it so fast. How? You know the old adage: it takes a village to raise a child—well it’s going to take a village to continue to raise the EDW too.

Think of it this way: what often happens with an EDW is the data architects, BI developers, and system/business analysts come to “own” the data in the EDW. They are seen as the subject matter experts for whatever datasets are in the EDW and they are expected to fully carry the burden of understanding the ins and outs of the data and ensuring data integrity. If “wrong” data ends up on a report, it’s the EDW staff that usually gets called to the carpet. But who should really “own” the data? The technologist responsible for making data reportable but whose exposure to the clinical practice that produced the data consists of a tour of the nursing unit that one time? Shouldn’t it be the clinical analyst who is meeting with someone in the unit several times a week and is immersed in the specific dataset in question non-stop?

That the EDW shouldn’t own data integrity might still be a novel idea in some places, but with all the new clinical datasets coming into the EDW these days, organizations are coming to realize that the technologists can’t possibly maintain a deep understanding of all the data content in the EDW. The data needs to be owned by people who work with that particular subject area every day and have an understanding of the clinical world that data comes from. The EDW team should function largely as the forklift operators in the organization, moving data from one place to another to be owned and easily accessed by people with the subject matter expertise to know what to do with the data and how to tell if it is accurate.

These folks with the subject matter expertise are often called data stewards or data managers. They own the metric or data in question and are supported in defining, validating, and maintaining the metric or data by the EDW team. Appropriately aligning resources helps the EDW team do more with less because they don’t have to own metrics or data for which they have little context. But we can go even further to make our EDW resources stretch: how about actually farming out some of the EDW development to those outside of the EDW team? That’s where it gets really interesting. In the next post we will dive deeper into that controversial idea.

Posted in

Kevin Campbell

I have over 20 years of experience in healthcare business intelligence and performance improvement, including developing enterprise data warehouses for large hospital and clinic systems. My work with other healthcare consulting firms and desire to help healthcare organizations leverage scarce resources through innovative approaches led me to co-found DTA; I believe we offer a unique value and perspective to organizations struggling with outcomes stagnation or other problems.

We’ve helped clients across the country accelerate toward value-based healthcare delivery.

Let us do the same for you.