The cost of evolving technology

A recent study has suggested that there has been increasing use of CT angiography in the diagnosis of pulmonary embolism:

A new study, published in the October 2004 American Journal of Roentgenology, shows that during a nine-month period in 1997 –1998, 81 patients underwent CT for suspected pulmonary embolism through the emergency department at University Hospitals of Cleveland. That number increased to 349 during the corresponding nine-month interval in 2002-2003, said J. David Prologo, MD, lead author of the study and chief resident in radiology, University Hospitals.

The nearly five-fold increase in the number of patients who underwent CT for pulmonary embolism did not reflect a proportionate increase in the number of patients with the disease, however. In fact, only about 6% of patients in 2002-2003 had findings of pulmonary embolism, Dr. Prologo said. That compares to about 25% in 1997-1998.

And why not? It is a non-invasive test that takes less than a minute to perform. I order it many times myself, especially when I work in the ER. Let’s look at the data.

UptoDate suggests that studies show a sensitivity between 53 and 87 percent, based on the methadology, and a specificity above 90 percent. However, with the lack of high-quality data, we come to this conclusion:

CT angiography shows promise as a noninvasive test for pulmonary embolism. Significant experience is required to optimally obtain and interpret images, however, and a negative result should be interpreted cautiously pending further data on the performance characteristics of the test.

With the technology of CT angiogram evolving, the study suggests that many findings uncovered are of unknown significance:

CT is allowing radiologists to see very small abnormalities that were not previously detectable. Often, this is resulting in the reconsideration of clinical issues, Dr. Prologo said. For example, clots in very small arteries in the lungs are now identifiable with CT. “It is not always clear, however, that these findings are responsible for the patient’s symptoms or are a harbinger of future disease. Therefore, given the risk of treatment side effects as well as the time and cost commitment necessary for therapy, the decision to treat some patients becomes a difficult one,” he said.

As with MRIs for breast cancer screening, evolving technology is leading to these “findings of unknown significance”. Cost analysis, including cost of working up possibly insignificant findings, certainly needs to be considered when bringing new studies and tests to the mainstream.