I recently posted a note in which I asserted that the rules engine being developed by Mayo was not particularly novel in that pathology professionals have been deploying similar rules on their LISs for decades (see: Rule-Based Decision Support Tools Available on LISs for Decades). Dr. Curtis Hanson, chief medical officer of the Mayo Medial Laboratories and also a good friend and colleague, responded with the following relevant and probing comment:
Appreciate your points below but I think you are missing some key points....We and others have obviously used rule capability within the LIS systems for years but they are "after the fact" rules - meaning that the clinician had already placed the order and if we had identified a problem order, then we had to call and get them to change the order etc. For example, AFP [as a] pregnancy marker (instead of as a tumor marker) ordered in a 65 year old lady - wrong test. We have to pick up the phone/or email/or EMR messaging and call to get the new test ordered or permission to change the test. This new tool is done at the point of order and allows the lab to get their rules in place at the time the clinician orders.
You are absolutely correct that these rules could be written directly into Cerner or EPIC at the point of order but reality is when we are out there in the real world marketplace, nobody has the time, money, resources, or expertise to do it. Most places have done nothing and some have gone through a lot of effort and implemented 5 to 10 "best practice alerts" into their systems. The labs are lower priority than implementing rules that meet national CMS quality standards or other measures that clinicians and CIMO's more easily see. I can also tell you that medical lab directors all over the country have been asking us for this kind of tool - both as a "waste model" but also as a movement into the value based world where they can more directly affect the ordering of complex tests that are poorly understood by clinician. It can be an educational tool at the point of order or it can be used as either a hard or soft stop tool. It's like all middleware systems - fill a gap that the big vendors (EHR or LIS) can't or won't fill.
Curt is correct with all of his points. Despite having rule developing capabilities for decades, most lab groups have used these capabilities to develop only a relatively small number of rules (what he calls "best practice alerts") that operated after-the-fact, which is to say after an order has been placed. This means that, as in the case of alerts about redundant testing or ordering the wrong or obsolete test, the clinician needed to be contacted and apprised of the problem. This was labor intensive and a task often relegated to house officers who would tactfully point out to the test-ordering clinician the problem and try to convince him or her to retract or correct the order. All of this occurred in a hospital culture where pathologists were often reluctant to try to sway a clinician's test ordering habits for the fear of engendering ill will. In the case of private pathology groups, an "irritating" pathologist could face job loss.
As to Curt's comment that the Mayo system is superior to LIS-based rules systems because the rules are fired real-time on the EHR side when the order is placed, he is also correct. However, it's useful to consider "why" rules deployed on the LIS can't fire in real-time for a test order placed on the EHR. Part of the problem is the speed of the network and the design of the interface between the EHR and the LIS. However, ponder for a moment that you can withdraw currency from an ATM in China working real-time with a large chain of computers that span the globe. On this basis, one can speculate that the LIS-EHR communication challenges in neighboring hospitals do not lie in the technical realm.
In short, EHRs and LISs could easily communicate in real-time in hospitals in the same way that EHRs in two hospitals in the same city but from different vendors could easily communicate to transfer clinical records (see: NYT Op-Ed on EHR Interoperability Blames Vendors and Greedy Hospitals; Hospitals Need to Pressure Their EHR Vendors to Improve Software Functionality). The reason that this task is difficult to achieve is that EHR vendors like Epic do not want to cede ground for the deployment of useful features to a competing EHR or LIS vendor. In others words, hospitals are denied useful functionality by their EHR vendor on the basis of the vendor's anti-competitive behavior. Specialized vendors like the one collaborating with Mayo, National Decision Support Company (NDSC), are allowed to participate in the Epic ecosystem because they play by Epic rules and are not big enough to threaten the Epic hegemony in healthcare IT.