The AMDIS Connection: The View Across Time & Patients

Clinical data repositories (CDR) have been the stepchild of the EHR for a long time. The CDR has traditionally been a research component that, after you dealt with billing and all the day-to-day information management of an EHR, you put together as a database that allowed you to look at the data across patients for planning, for financial reasons or research. But it was often done late.

At the University of Michigan, I was first exposed to this in a big way when we tried to get together a commercial EMR. We were trying to transition from one home-grown scheduling system to a commercial scheduling product for the ambulatory system, which handles for millions of patients at Ann Arbor.

We put together an Oracle database. We said, ‘let’s make it simple because it’s going to be temporary. We’ll do it in Javascript and we’ll make a web front end.’ While we were doing that, it wasn’t hard to download the information from the legacy system for scheduling into the web client and this relational database and then put a front end on it. Then someone said ‘Could you add the radiology results to it because it would be so wonderful to be able to access them in an easy-to-use, web-based way.’ Then someone else came around asking about laboratory results…

Within six months, we had CareWeb, a system that is still in place today—after more than 10 years. When the needs drive the system specific to an organization, it proves incredibly valuable. When you try to retrofit an assumption about a model and a presentation layer, you have to compromise. Vendors couldn’t possibly bring all your individual needs and considerations to bear in order to build a general purpose tool.

CareWeb presents information simply, and it allows the ability to look across not just medicine, but surgery, medicine, inpatient and outpatient. You can see information laid out chronologically and simply, and the data display was as flexible as HTML. And it was graphical. Point and click it, and you get what you’re looking for.

CDRs have provided the grounding for many projects and successful databases around the country. Because you’re dealing with a relational database, you can put anything you want in it and then compare one thing with the other. For example, we went to the cardiology division, in particular the catheterization lab. And by simply looking at the data the year after we looked at this, we recovered millions of additional dollars that were right there in front of us, but our process didn’t reveal it because we didn’t look at enough cases.

EHRs are targeted toward individual patients over time, by many different providers. But when you try to do things like quality improvement or safety patterns, the idea of having to look across patients over time becomes obvious and imperative, and you need a CDR relational environment to do it.

The CDR can be a power tool, but you need to build it on real business and clinical and operational needs. You need something that has a wide perspective and data view that individual stand-alone systems don’t have. Built carefully, with realistic expectations, a CDR can be best understood as a different way to view the information world of healthcare.