You wouldn't ordinarily consider a No. 2 pencil to be an essential tool of nuclear cardiology, but until June, physicians and technologists at Cedars-Sinai Medical Center in Los Angeles would have been helpless without one.
That's because for some 20 years they have been required to blacken hundreds of tiny ovals on Scantron-like bubble sheets for every patient imaged in the department. The odious and error-prone task was necessary to feed demographic and clinical information into a results-reporting database.
Early this summer, however, the department launched in-house software that automates the transfer of quantitative information on myocardial perfusion and function to the database from a computerized algorithm that draws those calculations from reconstructed image data.
It's all part of the first phase of a project at Cedars-Sinai to improve accuracy and speed report turnaround time. The project has been prompted by an increasingly competitive healthcare market and shaped by the knowledge that referring physicians, often cardiologists themselves, have other diagnostic methods at their disposal if nuclear cardiology proves sluggish.
"They have alternatives to nuclear. In many instances they have an echo right there in their office and they can get the results right away," said Guido Germano, Ph.D., who directs nuclear medicine physics at Cedars-Sinai. "We know that nuclear is better for a number of reasons, but if we take two days to get the information back to the physician, that is definitely not going to help us."
Before the information bridge was built, physicians had to manually transfer quantitative data from the nuclear medicine computer to the bubble sheets, just so it could be fed into a separate database that would automatically generate a report to the referring physician.
Documenting myocardial perfusion alone could involve filling in 100 bubbles. For example, in calculating the severity of perfusion defects, the myocardium is divided into 20 segments. Each of these segments is scored on five-point scale, depending on how abnormal perfusion is. This process is repeated as many as five times, for the rest, stress, and possibly, late-redistribution portions of the study, as well as for two separate measures of defect reversibility.
Documenting function can add another 80 numbers, as the 20 segments are scored for wall motion and wall thickening during rest and stress, in both gated and ungated exams.
"All of this information needs to be bubbled in just right, which is crazy, because the algorithm itself is a piece of software that operates on a digital image, and it generates digital data, so it makes no sense to go through the bubble sheet and introduce this analog phase. And yet, that's what everybody does," Germano said.
Nuclear cardiologists at Cedars don't do it anymore, and before much longer, technologists gathering information from patients before nuclear scanning won't do it either. By the end of the year, the department plans to launch browser-based software that will make it easier for technologists to document demographic information, clinical history, and electrocardiographic results at networked PC workstations throughout nuclear medicine.
Data entry will still be manual, but the bubble sheet will step aside in favor of the PC, after 20 years of service.
The Web-based software will make patient information accessible from any workstation in the department. More important, in both phases of the project, software programmers have developed ways to check the data for consistency and completeness to ensure that errors will be caught immediately before a report can be generated and subjected to several rounds of review and correction. Inconsistencies are highlighted in red and accompanied by a natural-language explanation of the error.
Catching errors early is already speeding report turnaround time, and once the second, Web-based phase of the project is complete, same-day delivery of results to referring physicians will finally be possible, according to Germano.
Phase two will also make it easier to modify the types of information collected as reporting needs change, something that has been difficult to do, given the rigid format of the bubble sheet.
Direct capture of data from resting and stress electrocardiographic equipment is on the list for next year. At some point, it may be possible to interface with the hospital or radiology information system to receive demographic information directly. That is a lower priority for now, however, as commercialization of the software will likely target physicians in office settings.
"What we're looking at is being able to satisfy even the physician in private practice who truly doesn't have access to a hospital information system and who needs to make the reporting faster and more accurate," Germano said.
ADAC demonstrated a prototype of the automatic reporting software at the annual Society of Nuclear Medicine meeting in June. Cedars-Sinai will work with other vendors as well, Germano said.
