Say cheese: Patient photos paired with radiography cuts wrong-patient errors

Taking a photo of a patient’s face at the same time a point-of-care portable radiograph is obtained can more than double the rate at which wrong-patient errors are caught by interpreting radiologists, according to a study published in the August issue of Academic Radiology.

The impact of wrong-patient events was underscored in a 2009 report from the Pennsylvania Patient Safety Authority, which said more than 30 percent of adverse events in radiology were related to wrong-patient errors, and nearly half of these occurred in the modality of radiography. Srini Tridandapani, PhD, MD, of Emory University School of Medicine in Atlanta, and colleagues indicated that the rate of these types of errors at their institution is around 0.01 percent; however, with more than one million exams performed each year, the absolute number of errors remains a cause for concern.

Thus, Tridandapani and colleagues sought to test whether pairing a photograph of a patient’s face with a portable chest radiography might be able to cut down on these types of errors. They recruited 90 radiologists to participate in the observer study. A total of 166 radiograph-photograph combinations were obtained from 30 patients, with consecutive radiographs from the same patients creating 83 unique pairs. Each radiologist interpreted a randomly chosen set of 10 radiographic pairs, containing up to 10 percent randomly generated mismatches. Radiologists were randomly assigned to interpret radiographs with or without the aid of patient photos.

“We found a significant improvement in their ability to detect wrong-patient chest radiographs when photographs were introduced, irrespective of whether they specialized in chest radiology or whether they had current expertise in interpreting adult portable chest radiographs,” wrote the authors. The proportion of errors detected increased from 31 percent to 77 percent, and the odds ratio for detection of error with photographs to detection without photographs was 7.3.

Results also showed that the presence of a patient photo did not significantly impact interpretation time.

“We suggest use of this identifier as an adjunct to, rather than a replacement for, the extrinsic identifiers required by the Joint Commission,” wrote Tridandapani and colleagues.