Skip to main content

Image Archiving and Reliability Managing Your Risks

Robert Petrocelli, MS, BS, CEO, Heartlab, Inc., Westerly, Rhode Island
September 2002
When I was younger and training as a physicist, I learned a lot about laws of nature, the rules that guide the way things work. However, there was one important law that was never taught: the law of unintended consequences. Understanding the unintended consequences of going filmless will help you plan and make better selections for your archiving system. One important unexpected consequence of going filmless is an increased expectation for access to the data that in turn cause an increased demand for system reliability. When the industry lived in the analog world and film was the primary medium of exchange for cath lab images, we always ran risks, such as having problems with film processors, film being fogged, and the risk of films being lost. These risks were fairly discrete and usually involved only one film canister and very few cases at a time, unless of course, there was a catastrophic event like a fire at an off-site storage facility. Today’s digital archiving systems eliminate many of the risks we faced with film, but that is not enough. There are new expectations, greater access requirements, and of course, there are still risks new risks that can impact dozens of cases or perhaps more, unless careful precautions are taken to mitigate those risks. A careless action that in the past would result in the loss of a film canister could now result in the loss of hundreds or more procedures. As you evaluate reliability risks for current systems and establish requirement criteria for the purchase of future systems, I recommend you ask four important questions. It is much less costly and troublesome to deal with these issues upfront rather than to correct them later. 1. Where is the potential for data loss? The greatest risk in any data image archiving system today is data loss. This problem most directly leads to issues like the potential for recatheterization of patients when surgery requires images, and the cancellation of surgeries when images are not available on a stat basis. When evaluating any system for purchase, require the prospective vendor(s) to provide a risk analysis table. This information is part of the FDA good manufacturing process and explains where the vendor believes the vulnerabilities for risk lie. These vulnerabilities are universal. The primary vulnerability for risk is located in what most vendors call the online storage system. The online storage system generally consists of hard drives, which reside in computer servers on a network. Hard drives are mechanical devices that use a magnetic recording technology that is extremely vulnerable to damage, degradation or erasure. There are a number of steps that may be taken to protect drives, but certainly the minimum precautions to take involve assuring some level of redundancy in a system. When a case is recorded, how long will that case reside in online storage before it is copied to another location? The optimal solution is to copy the case to a parallel storage system so that if the online storage system fails, the case will be immediately accessible on the fail-over system. Make sure the prospective vendor has identified all of the risk areas for data loss and can demonstrate what they are doing to minimize those risks. 2. How will data loss be prevented as time passes and cases are moved to what the industry calls the near line storage system (which could either be tape or DVD technology)? In the case of tape or DVD-RAM technology, both of which are erasable, decision-makers need to be vigilant because the drives that read the data back to us also have the potential to erase or modify that data after the fact. Therefore, there should be separate copies of tape or DVD-RAM available as an absolute minimum requirement for data integrity. In the case of DVD-R, this media is non-volatile; the recorded media must exactly represent the data in the online storage system. Verification methods used when data is copied from the online storage and the long-term archival system should be outlined and presented as part of the evaluation process, including data modified in any way between the online and near line storage system. Given the unique issues associated with the different types of archiving media, identify what each vendor is doing to protect the data in the archive. 3. How are redundant copies of data in the archive maintained to protect against data loss if there is an adverse event in a data center? Or, on a smaller scale, how will an exam be accessed if the network has been interrupted, the server is down, or the archive device is not functioning, and access to the data is still required (for example, cardiovascular surgeries still need to be supported?) Vendors should have a straightforward method of producing duplicate copies and producing a second database that is independent of the primary servers to index those copies. An institution must be certain that even if its basic infrastructure is not functioning, the images a physician needs will still be available. The increased expectation for uninterrupted access to the data must be addressed. Be sure you understand the full ramifications of each vendor’s approach to this critical aspect of system reliability. 4. How reliable is the archiving technology? This is very difficult to support without recording significantly large numbers of cases. The reliability data from manufacturers for media, recorders and disk drives by no means tells the entire story. What really needs to be understood is the reliability of the system in actual field conditions. This can be measured by the total number of cases recorded, divided by the number of cases that have been lost. An institution must record hundreds of thousands of cases, if not a million or more, before understanding the reliability of their system. Vendors can provide statistics on the overall reliability of their technology, When reviewing these statistics, keep in mind the size of the installed base volumes, which affects statistical significance. The final advice I give on risk management and digital archiving is to realize that even with the best due diligence and data protection standards, human beings ultimately manage the systems and are subject to errors which can possibly result in the loss of data sometime, somewhere. These losses usually do not result from failures in system design or hardware, but from failures to follow prescribed procedures. The risks can be mitigated by eliminating as many maintenance steps as possible that involve human interaction and by selecting a vendor with top-notch customer service to partner with your organization and provide expert technical support. Every institution should also have a risk management procedure in place so they can move quickly in ascertaining the nature of a data loss, formally documenting it, and notifying the affected clinicians who had patients in the affected time window. Address system reliability as a key aspect of your image archiving system. Ask questions that get to the heart of system reliability issues in your due diligence and be cautious of any vendor promising 100-percent reliability in a storage system.
NULL