Surveillance bias, the idea that “the more you look, the more you find” can lead to errors in outcome reporting, according an article in the Journal of the American Medical Association.
In the article, Peter Pronovost et al report that differences in outcomes may be related to surveillance bias rather than differences in quality. An example is that some clinicians screen at risk trauma patients for deep vein thrombosis using duplex ultrasound, while others do not screen asymptomatic patients. The screening leads to variability in the incidence of DVT reported.
Pronovost also notes that surveillance bias also has the potential to pose significant harms, for example, incentives to improve outcomes may encourage clinicians to avoid appropriate diagnostic testing to minimize reported complications.
Three steps are suggested to avoid surveillance bias:
- Those developing quality measures must clearly state surveillance methods;
- Policy makers look at cost and patient benefit in order to prioritize which measures to mandate;
- When surveillance is too expensive or risky, performance measures could link a process of care measure with adverse outcomes.