Friday, April 1, 2011

Tips from a Road Warrior (10): Keep Your Eyes Wide Open

One of my never-ending fears is that I forget something when going through the security check points in airports. I try to have a system for collecting my things after passing through the screen, i.e. first I retrieve my computer and other stuff, and then my bag and shoes, so I have less chance of leaving something behind. Apparently, this is not even a full-proof solution for everyone. Once I heard an airport announcement for someone at gate five to pick up his shoes. 

When I am fit and ready to leave for a trip, there is less chance of forgetting something. However, if I am suffering from severe jet lag, which makes me feel like I am sleepwalking, and I am transferring somewhere overseas, e.g. from London Heathrow to Gatwick airport, I am surely more prone to forget something. I actually only forgot my laptop once, which is not a bad record given the many trips I make, and fortunately I found out in time. I was checking through Munich after teaching a class and was half asleep because of the time difference. I was relaxing in the airport lounge, ready to go to the gate. Fortunately something triggered my consciousness when picking up my computer bag finding it to be very light. I rushed back to security and yes, they had stored my computer securely and after identifying myself I was able to retrieve it, and hurry to the gate. 

When dealing with medical information such as patient demographics and images, one cannot afford to suffer from jet lag, sleep deprivation or not being 100 percent fit. I know that it is hard if you get a call in the middle of the night to fix a study that was unidentified, rejected by the PACS, or not posted on a work list for whatever reason, but you have a responsibility to the patient to make sure you don't forget anything. The best way to do this is to have a fixed rule or process in place that is easy to follow and to check the fix when you are fresh and bright in the morning. 

When fixing things, also make sure that you use the right tools and know what you are doing. I often hear professionals using tools that they find for free on the Internet, which might not be validated or tested for use in a clinical environment. Now, there are many great free and open source utilities, many of them I use myself, but just make sure that when you use one of them, you ask around and/or do some testing with these prior to relying on them. Let me give you some examples of what I have seen in the field. 

There are devices that create duplicate unique identifiers (UIDs). By definition a UID is supposed to be unique worldwide and used to uniquely identify studies, series, images, frame of references, etc. It is used to index the database, for storing and retrieving. Therefore, duplicate UID’s used for different entities undermines the UID system and creates major problems. If a device has a poor UID generator, one might need to replace and fix this UID. One solution I heard of is that some administrators take the UID and add a ".1" to the end of the UID string. This is very dangerous as this is not necessarily unique. The solution is to use a tool that creates a new UID, using its own “root.” 

Another example is when one changes the image header with a utility that does not recalculate and update the number of bytes in a particular group. The problem is that early versions of DICOM had a so-called “group length” attribute in the header, which indicated how many bytes there are in a particular group such as the patient information group. These attributes have long been retired, and are rarely being created, however, some applications still create and/or use this information or just check the value, and, if incorrect, will reject the image or report an error. 

Another example is the correction of incorrect overlay information, which is not uncommon for ultrasound images. If a technologist forgets to change the patient information when scanning a new patient, it is possible that the incorrect patient information is “burned in” as an overlay into the image data. A similar, but even more severe patient safety issue occurs when the Left/Right marker is on the wrong side of an X-ray. In both cases, the pixel data has to be changed, assuming that an exam retake and/or recapture of the image is not possible. Many administrators simply use an overlay utility and put “XXXX” over the incorrect text or markers. This might appear correctly on a PACS viewing station, but when displaying on a different vendor workstation, teleradiology, or web-based viewers these overlays might disappear leaving an incorrectly identified image. This is also an issue when migrating the images to another PACS vendor system as overlays are often not migrated. One should use a tool that eliminates and replaces the actual pixel values. 

In conclusion, it is important to be alert and keep your eyes open and use the right tools when dealing with patient and image information. 


Many hospitals are connecting their outpatient facilities and clinics as well as their high-volume users of image and related information databases for exchange and access on a routine basis. Most use a secure Internet connection with a VPN to make sure that the patient privacy and security requirements are met. It gets a little bit more complicated when information has to be exchanged between organizations that are not part of the same delivery network, as patient information, especially Patient ID, often differ at the sites. Even more troublesome, the Accession Number, which is often used as a database key may be a duplicate of a number already in use. 

Some organizations have used tricks, such as automatic prefixes to the Accession Numbers that can be stripped as needed to prevent issues with data integrity. However, most organizations have a semi-automated import procedure for importing images from the outside. 

The good news is that the process of importing images through a VPN is virtually identical to importing images from off-line or exchange media such as CD's. Institutions have learned that it is very dangerous to import images from a patient without checking the patient demographics to make sure there are no data integrity issues. They typically create a new dummy order with a new accession number as well. In the case of images being transferred automatically, most PACS systems leave the images in the unverified or unspecified queue and allow a technologist or PACS administrator to check them and update any information as needed. 

It might be expected that these ad-hoc processes will be streamlined and automated in the US as soon as Health Information Exchanges or HIE’s come on-line. These HIE’s are currently in the process of being formed, mostly funded by federal grants, and will provide the infrastructure by which patient information as well as images can be exchanged between all participants. This will also provide the link with the National Health Information Network, or NHIN, so that information can be exchanged across all states. Critical to the implementation of these HIE’s are the Regional Extension Centers, or REC’s, which will provide technical consulting, vendor selection support for physicians to get their Electronic Health Record Software, and provider interface. 

The impact of the HIE rollout will be huge as, theoretically, every physician connected to the network will have access to any patient’s information, including images. It is almost guaranteed that the information exchange that currently takes place on a semi-ad-hoc basis and among regular high volume users over dedicated VPN’s will multiply as these gateways become the standard for information access. These HIE’s will facilitate the automated exchange and reconciliation of patient ID’s using standard protocols defined by IHE. 

There are still hurdles to overcome. Some states are moving very fast with their HIE implementations, and some are lagging behind. For example, at the recent Texas Health Information Technology Forum in Austin, Tony Gilman, CEO of the Texas Health Services Authority, showed how implementation in Texas is already well underway. State-level services are expected to take place in 2012, and the transition to sustainability will happen in 2013. 

The sustainability is still a risk factor for all of these HIE’s as they face long-term funding issues. Almost all HIE’s are initially funded by federal grants, for example, Texas received $28.8 million. However, after this money will be spent, who is going to support the organization and infrastructure? I am sure that this will take some negotiation between all stakeholders. 

In conclusion, image and information exchanges are increasingly shifting away from importing CD’s to the exchanging files over the Internet using secure VPN connections. These semi-ad-hoc connections will increase and, over the next few years, be replaced by more automated information exchanges using HIE’s whereby patient identifiers and other demographics will be automatically synchronized using the IHE profiles.