We have undertaken to review Bankart stabilisations performed by the senior author in Newcastle since 1998.

We employ a modification of the technique as described by Rowe et al, but without a coracoid osteotomy and using suture anchors. Where there is also an element of inferior instability this was combined with a capsular shift at the same time. We performed a retrospective case note review of all patients in the study period, following this all patients were sent out a postal questionnaire based on the Oxford Instability Score (OIS).

A total of 50 Bankart repairs were performed in the study period, no significant wound infections or haematomas were recorded. 3 patient had further dislocations and required revision surgery; a further 2 patients had ongoing symptoms of instability one of which has had revision.

Response rate to the questionnaire was 62%. Mean OIS for patients following primary stabilisation was 21.7 (possible scores from 12 – excellent outcome to 60 – poor outcome). OIS following Bankart stabilisation of the shoulder has not been previously reported. These results compare favourably to original scores published by Carr et al (1999) who included both patients treated by surgery and physiotherapy alone.

The OIS was shown to be very sensitive in detecting instability symptoms noted at clinical review. The response pattern of the scores closely mirrored that from Carr and co-workers.

The aim of the present study is to compare the Oxford Shoulder Score (OSS), the Shoulder Pain and Disability Index (SPADI) and the SF36 in terms of their correlation and sensitivity to change in a group of patients with subacromial impingement.

Patients attending a specialist shoulder clinic over a six-month period with a diagnosis of subacromial impingement completed the OSS, SPADI and SF36 on each of up to four attendances at six-week intervals, regardless of treatment. The data were analysed by creating scatter plots for each scoring system compared with each other, calculation of correlation coefficients, and calculating the effect size for each scale between subsequent visits (the difference in mean score divided by the standard deviation of the change in score for all individuals). This is a measure of sensitivity to change.

Three-hundred and twenty-three sets of scores were obtained in 110 patients. Scores were recorded across the whole range of the OSS and SPADI in our population. Similarly, the physical functioning component of the SF 36 ranged from 0 to 100. However, total SF 36 scores ranged only from 478 to 831.

There was good correlation between the OSS and SPADI scores (correlation coefficient = 0.86), although in 71% of observations the SPADI score was greater than the OSS when expressed as a percentage. The correlation between the OSS and SPADI was less good with the physical functioning component of the SF 36 (0.57 and 0.51 respectively), and worse with the total SF 36 (0.37 and 0.27 respectively) The OSS and SPADI showed good effect sizes (0.211 and 0.212) compared to the physical functioning and total SF 36 scores (0.104 and 0.063).

The OSS and SPADI show close correlation between each other, and were similarly sensitive to change in a group of patients with impingement. However, the SF 36 shows a poor spread of scores and less sensitivity to change in this group. These data support the use of the OSS or SPADI in these patients.

We assessed the inter- and intraobserver variation in classification systems for fractures of the distal humerus. Three orthopaedic trauma consultants, three trauma registrars and three consultant musculoskeletal radiologists independently classified 33 sets of radiographs of such fractures on two occasions, each using three separate systems.

For interobserver variation, the Riseborough and Radin system produced ‘moderate’ agreement (kappa = 0.513), but half of the fractures were not classifiable by this system. For the complete AO system, agreement was ‘fair’ (kappa = 0.343), but if only AO type and group or AO type alone was used, agreement improved to ‘moderate’ and ‘substantial’, respectively (kappa = 0.52 and 0.66). Agreement for the system of Jupiter and Mehne was ‘fair’ (kappa = 0.295). Similar levels of intraobserver variation were found.

Systems of classification are useful in decision-making and evaluation of outcome only if there is agreement and consistency among observers. Our study casts doubt on these aspects of the systems currently available for fractures of the distal humerus.