Radiographs of hip fractures were digitally altered to mask surgeons to the type of implant without compromising the reliability of quality ratings or making the rating process more difficult.

OBJECTIVE: To devise and test techniques to blind outcome assessors in trials of hip fracture fixation. STUDY DESIGN AND
SETTING: We developed three techniques (Blackout, Subtraction, and Overlay) to mask radiographs of hip fractures fixated with cancellous screws or dynamic hip screws. Fifty orthopedic trauma surgeons each assessed 32 radiographs blinded with each technique.
RESULTS: All techniques achieved low rates of correct identification of screw type (14.9% for Blackout, 26.9% for Subtraction, 22.1% for Overlay) and high proportions of "don't know" responses (72.3%, 48.4%, 52.8%, respectively). The interrater reliability of reduction quality in the blinded images (intraclass correlation coefficient [ICC]=0.55-0.57) was similar to the reliability of the unblinded radiographs (ICC=0.60). Surgeons perceived 6.9% of the Overlay images as much more difficult to rate than unblinded radiographs, compared with 9.7% of Subtraction images (P=0.25) and 28.0% of Blackout images (P<0.001).
CONCLUSION: Three techniques of blinding radiographs of femoral neck fractures successfully mask surgeons to the type of implant fixated, do not compromise reliability of reduction ratings, and do not make rating most radiographs more difficult. Trialists should explore creative approaches to optimize blinding when designing trials, and should incorporate rigorous approaches to testing blinding success.