In reply
We recently reported that approximately one-third of all imaging studies for pulmonary embolism (PE) are potentially avoidable based on a National Quality Forum endorsed performance measure.1 Inherent in the construction of any performance measure is the assumption that the care of patients can be simplified into necessary/avoidable, appropriate/inappropriate, good/bad. This simplification is readily apparent when we attempt to measure quality in the diagnosis of undifferentiated patients with potentially dangerous, nonspecific complaints. Dr Goodman is correct when he says that patients present for “signs/symptoms needing a rapid explanation,” rather than a “PE diagnosis.” However, we disagree with his tacit conclusion that any attempt to measure imaging appropriateness in complex patients is therefore futile. Rather than concluding that quality can only be measured at the level of an individual patient, we believe that results such as ours should be used to balance the potentially endless pursuit of diagnostic certainty against the risk that overtesting poses to our patients and the health care system.