Correlation found between two hip outcome measures

It may be possible to compare results between studies that use either the Oxford Hip Score or the Harris Hip Score.

At five-year follow-up, outcomes measured using the Oxford Hip Score correlate well with those measured using the Harris Hip Score, according to a study comparing the two systems in primary total hip replacement patients.

“This study implies that the physician-assessed and patient-assessed outcomes have significant correlation as would be expected in the medium term,” the authors said in the study.

“This good correlation also suggests that the case for the Oxford questionnaire, which does not require a clinical visit or physical evaluation, and had a higher follow-up rate, could be used more directly to compare studies that use the HHS (Harris Hip Score),” they noted.

Yegappan Kalairajah, MA, MPhil, FRCS(Orth), and colleagues at the East Surrey Hospital in Redhill, England, conducted the study. The data analysis included 115 hips of 115 patients who underwent unilateral THA as part of a larger study evaluating the Charnley Elite Plus (DePuy) prostheses. The researchers only included data for patients who had completed both HHS and OHS assessments at five years. All patients had HHS recorded preoperatively.

The Spearman correlation coefficient was used to compare outcomes between the two scoring systems. The researchers also used simple linear regression to report the regression equation and statistical values, according to the study.

At five years, patients had a mean OHS of 19.1, and their mean HHS had improved 50 points from preop to 89.4. The median OHS was 15 points and median HHS was 94 points, according to the study.

Using the Spearman correlation, Kalairajah and colleagues found good negative correlation between the two results (P<.0001), which was supported by linear regression.

Linear regression performed using the least squares method gave the equation Oxford = 109.22 - (1.039 X Harris), according to the study.

Using cumulative frequency distributions, Oxford scores suggested the following outcomes:

score of 12 to 18 - excellent;
score of 19 to 26 - good;
score of 27 to 33 - fair; and
score of 34 or higher - poor.

“The weighted k statistic is 0.679, suggesting good agreement in these categories, with the correlation strongest for those patients with excellent scores,” the authors said.

The authors noted one drawback to the study: No comparable OHS data were recorded preoperatively. However, “The authors feel that other authors have already addressed this preoperative-to-immediate-postoperative correlation and sensitivity conclusively,” they said in the study.

Other limitations included the use of only one type of prosthesis and evaluating only primary THA patients, they added.

For more information:
Kalairajah Y, Azurza K, Hulme C, et al. Health outcome measures in the evaluation of total hip arthroplasties - a comparison between the Harris Hip Score and the Oxford Hip Score. J Arthroplasty. 2005;20:1037-1041.