Patient Photograph Association With Radiologist Recommendations for Additional Imaging

Objective: Assess whether display of a patient photograph in the electronic health record (EHR) alongside head and neck CT or MRI radiology examinations is associated with recommendations for additional imaging (RAI) and whether self-reported race modifies that association.

Methods: This multi-institution health care system retrospective observational study from June 1, 2021 to May 31, 2022 included all patients with a head/neck CT or MRI report. We investigated association of photograph with RAIs using mixed-effects models adjusting for age, sex, complexity score, race, and area deprivation index while conditioning on patient and radiologist. Race was subsequently included as an interaction term. Multiple imputation was used as sensitivity analysis to address missing race data.

Results: In all, 60,543 reports were included from 48,143 patients (55.6% female; median age 58 years, interquartile range 40-70). The EHR included a photograph at the time 18.2% (11,048 of 60,543) of reports were signed. RAIs were included in 7.5% (4,522 of 60,543) of reports. Reports signed when a photograph was displayed had lower estimated odds of containing RAIs (odds ratio: 0.85, 95% confidence interval: 0.77-0.93, P < .001), consistent in sensitivity analysis, with no clear interaction between race and photograph (odds ratio: 0.99, 95% confidence interval: 0.69-1.46, P = .97).

Discussion: Patients with a photograph in the EHR had a lower probability of receiving RAIs and this difference did not seem to be the result of implicit racial bias but may be due to personalization of the encounter. This effect may influence radiology reporting for millions of patients per year. Further research is needed to determine whether the association has a positive or negative impact on care quality and outcomes.