Wednesday, June 1, 2011

Tips from a Road Warrior (12): Check Your Connections

The majority of my travel stories deal with my travels to India, as this country, despite its high tech image, still has so many infrastructure problems and issues that I typically characterize it as operating in "controlled chaos." Yes there is public transportation and there are taxis, and they have airports but they seem to be built for a fraction of the people that use it today. For example, there is typically only a single door to enter the airport, causing a queue out of the terminal stretching literally around the block. Unless you know a back-door, or wave a few hundred rupees, there is no way you will make it to the front without at least a one- or two-hour wait. As an illustration, I had to connect from Bangalore back to Dallas through New Delhi. Unfortunately, that means changing from the local to the international airport. Upon arriving at the local airport, there are no signs directing where to go or what to do, and after asking several people for a clue, they pointed me to an obscure ticket counter to purchase a bus ticket. The next bus was going to leave in an hour and looking outside to the queue already lined up, there was no way that we were going to fit into that bus. Talking with a fellow American in the same predicament, we decided to take a cab, which is an adventure by itself, especially when each of us had two suitcases to get into these tiny cars. Upon arriving at the international terminal, we decided to bribe our way through by giving an important looking person with an airport badge a good tip, and we made it just in time. 

Lesson learned is that any connection depends on the parties involved. Connecting from Tokyo Narita to Haneda or London Heathrow to Gatwick is a breeze, not in India. The same can be said about connecting devices, whether it is to a RIS or to a PACS or even a workstation. Just as traveling with the same airline, purchasing a modality from the same vendor as your RIS or PACS does not guarantee that it will be simple, straightforward, and without any connection issues. Remember most large companies have separate development centers for their different products. The RIS might be developed in the US on the East Coast, the modality in France, and the PACS by a division at the West Coast. Their DICOM conformance statements do not always follow the same template and do not have the same level of detail, and each seems to have their own conventions and ideas about configuration and installation. As a matter of fact, I would argue that some of the mid-sized PACS companies, who do not sell any of the modalities, are more open and easier to connect as they have every incentive to make this as painless as possible. 

It is important to do your homework and test the interfaces extensively prior to connecting devices. This not only applies to the interface of a modality with the RIS and PACS, but also to the reporting system, web-based viewers, MPR and future EHR applications as well. Most of the calls for assistance are not related to the basic connections anymore but for example, information missing in the report template for cardiology ultrasound exams. In this case, the reporting system does not quite interpret the information from the structured report template generated by the ultrasound. There are also issues with archiving and displaying the new modality objects such as being generated by the latest generation MR and CT as well as 3-D breast imaging devices. 

As images are going to be distributed more widely through electronic health records and possibly in personal health records as well, the connectivity issues are only going to grow, if for no other reason than rising volumes. Image distribution used to be rather well controlled within one or more departments, it is already widespread, so you can imagine what it will be like when patient images are accessible by any physician who is authorized by a patient: the sky is the limit. 

In conclusion, anytime there is a connection or interface, one should be aware of potential issues and prepare accordingly. This can be done by studying the interface specifications, and testing sample images and transactions in advance. Don't assume that it is easier if you use the same vendor on each side, just as using the same airline doesn’t guarantee easy connecting flights. 

Dose Issues Not Only For CT

There is a lot of activity around radiation dose reduction, especially for CT exams. This is partly due to the incidents that got a lot of press in which people were overdosed due to operator errors and negligence. Another important factor is the increase in CT exams, especially in the ER. It used to be that trauma cases would get a couple of X-rays done to look at potential fractures and/or internal damage, however, most ER's now have a resident CT and a body scan is pretty much standard procedure. 

After numerous studies raising safety concerns about the amount of radiation exposure for all these CT scans, vendors are finally taking notice and implementing techniques to reduce radiation exposure. One step being taken is to start registering the dose administered to the patient. This sometimes requires dosimeters in the X-ray chain, as well as reporting mechanisms. The reporting is still very much a work in progress. For some modalities, such as digital mammography, there already is relatively reliable information in the image header, which could be extracted by the PACS and stored. Some systems use the DICOM Modality Performed Procedure Step (MPPS) information as it also (optionally) can contain the dose. This is used in some cardiology applications whereby the cardiology information system records this information. There are drawbacks of the MPPS method as it is design dependent on the images that are created, and, for example, for fluoroscopy exams, there might not be any or only a few images taken. If one would depend on the dose information in the MPPS for those types of exams, the exposure would be severely under reported. 

This is especially true for CT, in which there is often a separate screen archived with the dose information, however, there is no digital representation, which means that the data extraction needs to depend on so-called screen-scraping or optical character recognition to get the actual information. The best way of reporting the dose information would be by using the dose-structured report. As a matter of fact, the Integrating the Healthcare Enterprise or IHE initiative has defined a special profile, called the radiation exposure monitoring or REM profile. This was demonstrated at the recent Radiological Society of North America meeting, however, there is a still a lack of recording and reporting systems, which is causing very slow implementation. 

CT dose reporting is getting most of the attention, however, in my opinion, the over exposure and unnecessary exams using standard X-rays such as DR or CR is underestimated. As an example, my little six-year-old grandson has issues with allergies and congestion, especially during the flu season. He has already had several chest X-rays over the course of his first six years as pediatric physicians like to play it safe and order an X-ray "just in case." There are also no guidelines on how much to reduce the technique factors to maintain reasonable image quality, and still to be able to make a diagnosis. The Image Gently campaign has developed online teaching materials, but to my knowledge, there are no guidelines published yet. In addition, if images are taken, there is also often a lack of shielding, an issue which was reported in the article “X-rays and Unshielded Infants” on Feb. 27, 2011 in the New York Times. 

It might seem strange to hear a message of reduction of X-rays from someone like myself who works in this industry, however, one should realize that 80 percent of the world’s population has no access to X-rays at all. Rather than over utilizing these systems for the privileged 20 percent, it might be better to expand it to those who have no access. This requires the development of low cost digital systems, which are very durable and easy to use. I believe that this can be done, if some of the major vendors would just make this a priority. 

In conclusion, dose registration is still challenging, the IHE REM profile implementation should be a major push. However, registration is just the first step, further development of dose reduction techniques and guidelines by professional organizations are needed as well.