Impact of Cerner’s CPOE study about increased mortality will likely be more regulation
Yesterday I spent some time being interviewed by reporters about the impact of the CPOE study (about Cerner) that I reported on a few days earlier. One of the reporters asked a great question: what will be the impact of this negative study?
My answer was two fold. One was that I didn’t see any big impact on Cerner itself because any objective reading of the Pediatrics article would see a few flaws in the study (which I also elaborated on at length). And, almost none of the HIT vendors are liked signficantly more than others. I said I didn’t see any wholesale jumping from one vendor (Cerner) to another because most people believe choosing a HIT vendor is like choosing a President — a choice is made for the one we dislike less, not love more. So, from a competitive standpoint I don’t see much impact. Cerner was not completely to blame nor were the users. It was a joint selection and a joint problem between vendor and customer. One of the things that probably will impact Cerner negatively would be if its solutation was unsuitable for the specific use at the customer (in an ICU) and they didn’t say anything during the sales/deployment/evaluation/test process.
The second part of my answer was that the biggest impact could come in the form of regulations. If the FDA were paying attention, they would see that lots of hospital information systems, especially those making it over to the clinical side through clinical decision support, should really be regulated. After all, they already regulate blood banking software and that software (which I’ve worked on for years) is not significantly “more important” or “more deadly” than other healthcare software like CPOEs. Without regulations we in the clinical informatics world do not spend the time or effort necessary to test for fitness or suitability for particular uses. We don’t perform large-scale simulations. We often don’t conduct risk analyses the way we should. In high-risk software like clinical systems, over 50% of the time spent in the software should be for design and testing (bookends of the development process). And, the design and testing should be well documented and materials publicly available if appropriate.
As a developer, enterprise architect, and entrepreneur who’s worked in the health IT world for a long time, I’d hate to have more government regulation in our industry than is absolutely necessary. However, if we’re not careful and we don’t do a good job policing ourselves, we won’t have anyone else but ourselves to blame when the FDA comes knocking on our doors wondering about how we develop high risk software.
I wrote a brief assessment of FDA’s power to regulate last month that you might find interesting as well.