Effect of training level on accuracy of digital templating in primary total hip and knee arthroplasty

The use of digital radiography and templating software continues to become more prevalent in orthopedics as the number of total hip arthroplasty (THA) and total knee arthroplasty (TKA) procedures increases every year. The purpose of this study was to evaluate the effect of training level on the accuracy of digital templating for primary THA and TKA. Digital radiographs of 97 patients undergoing primary THA (49 cases) and TKA (48 cases) were retrospectively templated using OrthoView digital planning software (OrthoView LLC, Jacksonville, Florida). Anteroposterior hip and lateral knee radiographs were digitally templated and compared with the actual size of the implants used intraoperatively. An implant sales representative, physician assistant, medical student (J.D.K.), resident (A.R.H.), and fellowship-trained arthroplasty surgeon (B.R.L.) templated all cases independently after a standardized orientation and were blinded to the actual component sizes used for surgery. The medical student, resident, and arthroplasty surgeon retemplated the same 97 cases 1 month later to determine intraobserver reliability. Digital templating was accurate in predicting the correct implant size in 33% of THAs and 54% of TKAs. In 73% of THAs and 92% of TKAs, digital templating was within 1 size of the actual implant used, and in 88% of THAs and 99% of TKAs, templating was within 2 sizes of the final components. In no cases did the templated implant size vary by >3 sizes from the final components. Interobserver reliability for templating THAs and TKAs showed good reliability as measured by intraclass correlation coefficient (ICC) (ICC(THA)=.70; ICC(TKA)=.86). Intraobserver reliability for templating THAs had excellent reliability for the resident and arthroplasty surgeon, with a kappa coefficient (κ) of 0.92, and good reliability for the medical student (κ=0.78). Intraobserver reliability for templating TKAs showed excellent reliability among all examiners (κ=0.90).