We studied 19 videotaped knee arthroscopies in 19 patients with mild to moderate osteoarthritis (OA) of the knee in order to compare the intraobserver and interobserver reliability and the patterns of disagreement between four orthopaedic surgeons. The classifications of OA of Collins, Outerbridge and the French Society of Arthroscopy were used. Intraobserver and interobserver agreements using kappa measures were 0.42 to 0.66 and 0.43 to 0.49, respectively. Only 6% to 8% of paired intraobserver classifications differed by more than one category. Observer-specific disagreement was evident both within and between observers. A small, but significant, occasional variation was also seen. Although reliability may improve by an analysis of disagreement, it appears that the arthroscopic grading of early osteoarthritic lesions is inexact.

In a prospective, randomised, multicentre study, 112 patients (99 men and 13 women, aged between 21 and 63 years) with acute, complete rupture of tendo Achillis were allocated either to surgical treatment followed by early functional rehabilitation, using a brace, or to non-surgical treatment, with plaster splintage for eight weeks. The period of follow-up was for two years. Evaluation was undertaken by independent observers and comprised interviews, clinical measurements, isokinetic muscle performance tests, heel-raise tests and an overall outcome score.

The rate of rerupture was 20.8% after non-surgical and 1.7% after surgical treatment (p < 0.001). Surgical and non-surgical treatment produced equally good functional results if complications were avoided. However, the rate of rerupture after non-surgical treatment was unacceptably high.