No-shows in radiology can be predicted—no crystal ball required

Twitter icon
Facebook icon
LinkedIn icon
e-mail icon
Google icon
 - calendar

No-show visits (NSVs) are a considerable obstacle for all healthcare specialties, and radiology is no exception. Imaging leaders have often wished they could predict which patients might be NSVs, and according to a new study published in the Journal of the American College of Radiology, that wish has come true.

H. Benjamin Harvey, MD, JD, department of radiology at Massachusetts General Hospital in Boston, and colleagues studied data from more than 54,000 scheduled patient appointments with radiology examinations at a large academic medical center. All appointments were scheduled from Jan. 1, 2016 to April 1, 2016. The team’s retrospective study revealed that information available in the electronic medical record (EMR) can be instrumental in successfully predicting NSVs in radiology; you just have to know what you’re looking for.

Overall, the data revealed a no-show rate of 6.5 percent. For mammography, that rate was more than 10 percent, while it was more than 7 percent for CT, more than 4 percent for MRI and more than 1 percent for PET.

The researchers also performed a regression analysis, which revealed that a patient’s prior no-show rate, the modality type, and whether an exam was scheduled two days in advance were the strongest predictors of a NSV taking place.

A patient’s employment status, whether they speak English as a primary language and the approximate distance from the patient to the health system were found to not be significant predictors.

Harvey et al. noted that this research provides imaging groups with valuable insight when it comes to scheduling future appointments.

“Our study has practical implications for radiology practices and health systems interested in improving clinical efficiency and optimizing patient outcomes,” the authors wrote. “First of all, NSV prediction model can be used directly, to develop dynamic scheduling and overbooking strategies based on predicted no-show probability. Likewise, waiting times for late patients could be set to a lower threshold for patients with a higher risk of no-show.”

The team added that this data can also help leaders implement specific interventions—extra text messages, for instance—to help reduce the chances of NSVs. “Accurate predictions of radiology no-shows hold promise for identifying patients who might benefit from additional patient engagement efforts with the aim of optimizing patient outcomes, improving clinical efficiency, and recapturing lost revenue,” they wrote.

The team added that their study did have limitations, including the fact that its data was from a single academic center. In addition, they wrote, “potential personal, environmental, and clinical reasons that could have contributed to a patient no-show but were not readily available in the EMR” could not be accounted for properly.