Validity of Radiographic Analyses Between Hand-Drawn and Computer- Aided Measurements: A Double-Blinded Test-Retest...

Spinal radiographic analyses are valid and reliable practices used for patient management in healthcare. Technologies and tools used for these analyses need to be valid and reliable.

This study investigates repeated accuracy for validity and intra- and inter-examiner reliability of computer-aided lateral spinal radiograph measurements using PostureRay^® EMR software.

Rectangle renderings representing lateral spinal radiographic views were created to remove examiner interpretation of anatomical landmarks as performed on actual radiographs to examine the objective repeated accuracy of PostureRay^® software. A blind, randomized, repeated measure design was performed using four examiners trained and experienced with the Harrison Posterior Tangent method of measurement. Each examiner analyzed the spine renderings using a hand-drawn and PostureRay ^® method multiple times. A fully crossed design whereby examiners performed multiple assessments per rendering using each analysis method was used to assess intra- and inter-examiner reliability within each method and a formal comparison of the two methods.

Using hand-drawn and PostureRay^® methods, examiners displayed very high intra-examiner reliability with correlation values greater than 0.999 for each combination of trials. PostureRay^® method showed consistent accuracy of measurements. Further, there was no statistical difference in average concordance between hand-drawn and PostureRay^® methods for the lateral cervical, thoracic, and thoracolumbosacral spinal regions.

This study reports a very high degree of intra- and inter-examiner reliability of radiographic line drawing methods and establishes concurrent validity of PostureRay^® EMR software in determining angles and displacements of lateral spinal alignment as an equivalent method to the handdrawn method.