Personal Telemedicine
Evan Rosen

Stored video is rapidly becoming critical to many telemedicine applications. The problem has been the lack of a way to catalogue and manipulate video information…until now. The solution comes from one of the 70 companies exhibiting at the DVC (Desktop Video Communications) '98 Spring conference and exhibition in Santa Clara, California. As I walked the aisles of the show, I ran into many of the usual suspects in the video communications marketplace, the companies whose names regularly fill these pages. Many of these firms are delivering exciting videoconferencing technology, but I was looking for something new that might have an order-of-magnitude impact on personal telemedicine. After lots of looking, plenty of listening to sales pitches, and restraining more than a few yawns, the cocktail hour came. I grabbed a beer and continued my quest. As I rounded a bend near the hors d'oevres table, there it was….

This was no fancy exhibit, but there was nevertheless some volcanic activity. Digital Lava was crammed into a tiny booth and Tom Stigler, the company's Vice President of Marketing, was conducting demonstrations. The company's name was familiar, because of a pre-show, surround-sound style public relations blitz orchestrated by a San Francisco PR firm. It all seemed like pure hype. What I found at the booth, however, was fascinating enabling technology to access, manage and publish video. As I downed my Heineken and watched Stigler's demo, I had a sense of Aha!

Digital Lava's product line includes vPrism, a software-only video publishing system that lets users organize video content, link the video to other types of files, and publish the video with all of the linked information as VideoCapsule files. VideoVisor is the client software that allows access to VideoCapsule files, analysis, re-arranging of content, annotation, subtitling of text and transcripts, and the ability to view notes and links.

Echocardiograms: frame-by-frame analysis

"If you've collected medical information on video like an echocardiogram and the person doing the diagnosis sees irregularities in heart beat, they could mark the point, make notes and put a link to an image or document in a patient's record," says Stigler. "Related files can be linked to video at the frame level." Stigler points out that traditionally a physician's written diagnosis can be linked to a video clip, but not inside the clip. Digital Lava's system lets a doctor make notes on a single frame of video and then link the frame to the notes and other related images and Web sites. "Previously video has been a globby, voluminous file. We're making it manageable," Stigler insists.

Digital Lava ( is the brainchild of Dr. Jim Stigler, a U.C.L.A Professor of Psychology and brother of Tom. In 1993, Jim Stigler was conducting a cross-cultural education study in Japan, Germany and the U.S. As part of the study, he was analyzing differences in non-verbal communication. His team had shot 300 hours of video, which Jim Stigler wanted to analyze frame-by-frame. So, Jim Stigler brought in a software development team to create video analysis tools. He then founded Digital Lava in 1995.

Digital Lava could become to the burgeoning field of video publishing what Adobe has become to desktop publishing. The company has inked a technology licensing and reselling deal with RealNetworks. Also, Digital Lava has signed up b4 Technology in Bletchley, UK as a distributor in Europe. The vPrism publishing system runs on Macintosh, Windows 95 and Windows NT. The server software costs $8995. The client software costs $268, with volume discounts available.

Intel's Secretive Lab

Digital Lava is not alone in its quest for the video publishing market. At Intel Corporation's secretive Microcomputer Research Laboratory in California, the husband-wife research team of Boon-Lock Yeo and Minerva Yeung are working on ways to make stored video and still images more accessible. The tools that Yeo and Yeung are developing let the user sift through hours of video and quickly understand the content. The tools then create a visual summary of the content, which would be represented as an image or a composite of a few images surrounding a primary image.

"The use of video today is based on linear, sequential access," explains Yeo. "Our approach is to break videos into small, atomic units--scenes, shots--in an intelligent fashion. What you then have is random access capability." Yeo likes to use the analogy of a reader's ability to flip through a book. His software tools allow jumping to any point in the video. Because there is much repetition in most video content, Yeo and Yeung's research is more focused on shots than on frames. For example, Yeo points out that a half-hour newscast includes between 200 and 300 shots. The team's tools let the user display thumbnail representations of these shots on screen simultaneously. While television and advertising are among the primary applications that Yeo and Yeung have been investigating, Yeo believes that telemedicine might also be a good match. "In telemedicine, you would want some form of browsing, plus searching capabilities," says Yeo.

Yeo and Yeung, who both hold PhD's in electrical engineering from Princeton, developed an interest in stored video tools while in graduate school in 1993. Soon after, they began their research. Next came a 2-year stint at IBM's T.J. Watson Research Center in New York, where the fruits of some of their efforts became part of DB2 Universal Database, release 5, IBM's relational database product. Yeo and Yeung then joined Intel's 3-year-old Microcomputer Research Laboratory (MRL) in January of 1998. Now the team is talking with potential licensees of the technology they have developed at Intel.

Killer apps for store-and-forward

While still images are used widely in telemedicine for exchanging x-rays, MRI's and CAT scans, stored video has more specific applications. These include capturing echocardiograms in telecardiology and fetal video in telegynecology. One software company, Second Opinion, is planning an August release of a stored video enhancement to its medical imaging and data collaboration product. The product gives users the option of forwarding the stored video or playing the video during a videoconference. The second option avoids problems with large file transfers over low-bandwidth connections. Second Opinion has already received an order from New Brunswick Heart Center in Canada, which is using MPEG-2 compression over regular phone lines. Several other medical technology companies are also planning store-and-forward video products. One such company says that its market research shows that telecardiology is the killer application for store-and-forward video in medicine. This is because the technology requires little change in habits on the part of cardiologists and because the doctor receives reimbursement regardless of the recording and playback media.

At Oregon Health Sciences University, Dr. David Sahn leads a team of pediatric cardiologists who are conducting real-time telemedicine consults. Sahn is a pioneer in the field who did the first 2-D echocardiogram on a baby in 1971 while a fellow at the University of California-San Diego. Dr. Sahn and his team use Picturetel systems running at 384 kbps to guide medical technicians and physicians while they are doing the echo. "The only way to do cardiac images is with moving images, because the heart moves," according to Dr. Sahn. "In contrast, to view a radiology image of a liver or a kidney, a still image is fine."

While OHSU is currently using real-time video, Dr. Sahn believes that store-and-forward will replace some of the real-time consults. "The major disadvantage of real time is that both people have to be there," he notes. Nevertheless, pediatric cardiology presents a specific set of problems that must be overcome before store-and-forward digital video can be widely adopted. Pediatric cardiology requires that the medical technician perform the echo to more specific specifications than adult cardiology requires. Therefore, pediatric cardiologists prefer to instruct the technician as the procedure is performed. "You can't take a technician who doesn't know what he's looking at and have him capture the video," explains Dr. Sahn. "Tight protocols have to be established."

The DICOM international standard

Once the protocols are in place for store-and-forward pediatric cardiology, Dr. Sahn believes that stored clips can become a useful part of a medical record. He also says that the ability to annotate and sift through these stored images will become important. "The first choice would still be real-time. But if the specialist is not there when the echo is happening, the choice would be store-and-forward with some notes on the frames." Dr. Sahn adds that ultimately store-and-forward telemedicine should comply with the Digital Imaging and Communications in Medicine (DICOM) standard. However, many medical imaging vendors have been slow to implement DICOM, a de facto standard developed by the National Electrical Manufacturers Association and the American College of Radiologists in 1993. This is partly because the standards open a customer's doors to products from other vendors and allow users to mix and match. This in turn threatens to cannibalize profits. Mark Oskin, a post-graduate researcher at the University of California-Davis who is implementing DICOM there, believes DICOM is critical to multi-vendor situations. Nevertheless, "It's been a slow momentum. It's not easy to implement and there's a lot of money [for vendors] in remaining proprietary,"according to Oskin.

David Balch, Director of Telemedicine for East Carolina University's School of Medicine, confirms the killer applications. "Echocardiograms and OB-GYN studies. That's where all the action will be in the next 12-18 months for store-and-forward," says Balch. "There will be a big splash in that." Balch believes that better-integrated tools are needed for stored video in telemedicine. "You need a better user interface, fewer clicks to get it done, and as volume builds you'll need a central way to do storage and recall." Nevertheless, Balch is less certain that tools allowing doctors to annotate frames of video will be used widely. "They get a video tape, they play it, they write a report and they're done. They won't want to do anything that takes any more work," insists Balch. Still, as a new generation of physicians becomes more technologically-astute, the demand may grow for frame-by-frame video analysis and annotation tools. The most thorough specialists could find that such tools make a complicated diagnosis easier.

Jim Bruton is among the biggest troopers in telemedicine. The Emmy-award winning documentary producer was the technical brains behind the now-famous Mt. Everest telemedicine project that used Zydacron technology. In May, Bruton and a team of medical experts hiked for more than seven days to the Mt. Everest base camp at 17,500 feet. There they set up a telemedicine clinic with a satellite transport that allowed doctors at Yale Medical School to evaluate climbers and the native Sherpas. As part of the project, Everest Extreme Expedition climbing team leader Wally Berg wore a two-pound BioPack with remote sensors designed by MIT. The objective was to transmit vital signs and medical data to medical personnel at M.I.T via satellite (see related story, p. ____).

I kept in regular e-mail contact with Jim Bruton while he was in base camp. "There's no manual written for a lot of what we're doing," according to Bruton. "When you're walking into virgin territory, you're going to learn something." While the telemedicine clinic at base camp succeeded, the M.I.T-designed BioPack did not fare as well. Mike Hawley, the team leader with M.I.T, confirms that there were problems both with transmission and batteries. According to Hawley, "The upshot was that we could not be confident the packs were functioning well past Camp III, decided that for a first step we had collected enough data, and instructed the climbers to remove them." As for Jim Bruton…he is safely back in Connecticut and has a guest gig as a lecturer in anatomy and experimental surgery at Yale.