Several days ago, I wrote a piece griping about the vague use of the term “cloud” in describing networking strategy — and particularly, the notion of the cloud as platform upon which to deploy an EMR. My gripes didn’t even take on the idea of whether virtualized EMRs were a smart approach; I was just troubled by the way vendors were throwing the term around, and perhaps selling vastly different products under the same moniker.
Well, I give, at least for the moment. The more I read about cloud deployments of EMRs, the more it appears that something useful is going on. (Cloud security is still a question mark, but that’s a subject for a whole ‘nother article.)
If nothing else, cloud-based models of EMR deployment seem to be easier and more flexible to use than some SaaS models, as client users almost certainly won’t need to download additional apps or even browser scripts to use them.
Some recent examples of cloud deployments in the hospital setting:
* This month, a pair of London hospitals began storing patient data in the cloud. The two hospitals, Chelsea and Westminster Hospital, are part of a National Health Service pilot known as E-Health Cloud. Since writing this item up in July, I’ve learned that the private cloud effort will be supported by a Scottish cloud vendor named Flexiant. To fend off worries over security, users will have to pass multiple ID checkpoints to get to patient data.
* Hudson, NY-based Columbia Memorial Hospital has maintained a private cloud-based platform to support its eClinicalWorks system since early 2010, one which serves 26 clinical locations and 300 providers. Hospital CIO Cathleen Crowley told SearchHealthIT that the cloud approach has allowed the facility to use less physical servers and minimized the hospital’s need for IT support for the EHR. (Interestingly, the hospital is also participating in a pilot HIE project, bolstered by a $1.03 million grant from the state’s Department of Health.)
* Harvard Medical School and Beth Israel Deaconess Medical Center have taken what leaders see as the best of the public and private cloud models to provision their EHR. As defined by then-CIO John Halamkha, the public cloud involves rapid provisioning of CPU cycles, software licenses and storage, but no guaranteed service level or strong security. To address these limitations, the two institutions built Orchestra, a 6,000-core blade-based supercomputer, designed to be highly secure and available, as well as adding grid technologies to share CPU cycles among high performing computing facilities nationwide. This superplatform offers a virtualized environment for 150 clinical offices, hosting 20 instances of logically isolated EHR apps per physical CPU. (By the way, I believe these institutions use a home-grown EHR of their own.)
What can we take from these stories? Well, we don’t have enough information to draw scientific conclusions, but it seems that a) building out private cloud virtualization of EHRs can be (very) expensive, but ultimately works well and that b) Securing data in the cloud still takes a long ton of effort.
The next challenge for the vendor community, it seems clear, is to offer a cheaper private cloud infrastructure that hospitals trust. Not sure how long that will take, though. Heaven knows this will prove to be a big challenge.