I have posted a number of previous notes about the computer technology that goes by a number of names including ASP architecture, SaaS, and cloud computing. Simply put, these terms refer to an architecture whereby access to servers, application software, and data storage are, in essence, rented from a specialized supplier of these services.
There are a number of advantages to such an approach. For example, when a healthcare software vendor like Cerner provides LIS remote hosting for its client hospital labs, it also provides a deep knowledge about optimization of the system. Cloud computing also offers financial incentives since access to remote servers is now a commodity and much less expensive than having a lab or hospital maintain its own local computer room with hardware.
A recent note in HIStalk by a reader, Stratto Cumulus, provided a number of reasons why cloud computing in healthcare can potentially cause problems. Below is the note in its entirety (boldface emphasis mine):
“Re: cloud computing. Reminds me of the late...unlamented ASP model, where clients wanted to outsource everything, lay off staff, and make huge profits. In real life, the vendors were running cloud servers with vanilla COTS [commercial, off-the-shelf] applications that they would not modify, which killed the business since enterprise apps always need customization and interfacing. Questions to ask: who’s handling authentication and security? Will the cloud vendor tell you if someone snoops in George Clooney’s records, or if you suspect someone is, how fast will they look into it? How will COTS licensing be handled, or will only open source stuff be in the clouds? Will the cloud be a data repository with local data marts or will local systems collect the data and batch it up to the cloud? Will it be transaction driven and Web-based, and if so, how many critical clinical apps are really Web-enabled? How will APIs and web services be handled? Are you sure you have the bandwidth? Clouds could be great for research repositories, provided authentication and architecture is adequate to handle the multiple query services. It would be great if we could integrate research findings across multiple studies to increase statistical power or see relationships across organizations, genes, etc.
If one is anticipating the use of cloud computing for LIS or EMR support, it would be wise to keep in mind all of the questions posed by Stratto Cumulus in his note and raise them as part of the due diligence or the RFP associated with the project. This is probably an unlikely event because this type of architecture has not been widely adopted for healthcare IT products. However, in a previous note (see: A Closer Look at the Vertical Cloud in Healthcare Computing), I did discuss how remote file storage is being used for PACS file storage.
Comments