Intra- and inter-rater reliability in photogrammetric pelvic tilt angles analysis

Abstract

Background/aims: Digital images are an inexpensive and complete method for registering pelvic positions and their changes as a result of clinical intervention. This study aimed to evaluate intra- and inter-observer reliability of pelvic tilt angles from anterior, posterior and bilateral views. Methods: Photographs of the anterior, lateral and posterior views of subjects (n=93) were analysed by Alcimage software, which marked pelvic tilt angles (right (R) and left (L) side views): between the anterior superior iliac spine (ASIS) and posterior superior iliac spine (PSIS) and one horizontal line; beyond the angle between the two ASIS and one vertical line (ANT); and between the two PSIS and another vertical line (POST). Three observers evaluated the photos twice. Intra- and inter-observer reliability were assessed using the t-test, correlation coefficient, coefficient of variation (CV) and intraclass correlation coefficient (ICC) (P<0.05). Findings: The intra- and inter-observer t-test showed no significant difference. The intra-and inter-observer correlation coefficients varied between 0.9994 and 0.9760 and between 0.9510 and 0.9941 respectively. The ICC and CV between intra-observer measures were classified as excellent for the angles R and L but not for the angles ANT and POST. Inter-observer ICC was also high for the R, L and POST angles but poor for the ANT angle. Conclusion: The angles analysed from a sagittal view using photogrammetry enable the monitoring of possible longitudinal changes in pelvic tilt. However, the angles analysed from a frontal view may not be suitable for the diagnosis of pelvic asymmetries.