Introduction
A multitude of hip function scores are used to report outcomes in clinical studies, which tends to restrict the comparison of clinical outcomes. The present study attempts to provide an equating relationship between Harris hip score, WOMAC Index and Oxford hip score.
Methods
75 patients attending the hip arthroplasty clinic were asked to complete the Harris hip score (HHS), Oxford hip score (OHS) and WOMAC Index (WOMAC). Rasch analysis was used for comparative analysis of all three scores. Initially, the Harris hip score was used as a reference frame to generate person measures. The WOMAC Index function score and Oxford hip score were subsequently added to the analysis. Measures from this analysis were used to generate a comparative equation for three scores. Comparative equations from above analysis were used to calculate WOMAC Index and Oxford hip score using Harris hip score a baseline. The calculated scores were used to construct a nomogram. WINSTEPS ® software was used for analysis.
Results
Equating Harris hip score and WOMAC Index resulted in the following equation -:
WOMAC Index = 2.84 - Harris hip score*0.432. Both instruments had a correlation of -0.43 and disattenuated correlation of -0.59. According to above indicators there was some degree of measurement error. Analysis of Oxford hip score with Harris hip score as a reference frame resulted in the following equation-: Oxford hip score = 0.411+Harris hip score*0.793. This analysis showed a higher correlation of 0.79 and the disattenuated correlation was 1.0. The scale units for nomogram were adjusted according to the difference in variance of each score.
Conclusion
The equations from analysis reflect the scoring systems of individual scores. The numeric functional unit of Harris hip score and WOMAC Index is inverse whereas Oxford hip score and Harris hip score are direct. This is similar to the routine clinical assessment and substantive for the above analysis. This study is the first attempt to compare and equate hip function scores. The results can be helpful for a quick clinical comparison and for comparing large studies using different outcomes. This study (N=75) is an initial attempt to compare functional scores and an increase in sample size will result in more robust assessment and generalization.