The Shoulder Function Index (SFInX): a clinician-observed outcome measure for people with a proximal humeral fracture

BACKGROUND Proximal humeral fractures are amongst the most common fractures. Functional recovery is often slow and many people have ongoing disability during activities of daily life. Unidimensional measurement of activity limitations is required to monitor functional progress during rehabilitation. However, current shoulder measures are multidimensional incorporating constructs such as activities, range of motion and pain into a single scale. Psychometric information of these measures is scarce in this population, and indicate measurement issues with reliability. Therefore, the aim was to develop the clinician-observed Shoulder Function Index (SFInX), a unidimensional, interval-level measure of 'shoulder function' based on actual performance of activities, reflecting activity limitations following a proximal humeral fracture. METHODS An outcome measure development study was performed including item generation (existing shoulder measures, focus groups) and item selection (selection criteria, importance and feasibility ratings, pilot testing, Rasch analysis). Clinicians (n=15) and people with a proximal humeral fracture (n=13) participated in focus groups. Items were pilot tested (n=12 patients) and validated in a Rasch study. The validation study sample (n=92, 86% female) were recruited between 5 and 52 weeks post-fracture and had a mean age of 63.5 years (SD13.9). Measurements at recruitment and 6 and 7 weeks later were taken in three public metropolitan hospitals or during home visits. Raw SFInX data were analysed with WINSTEPS v3.74 using polytomous Rasch models. RESULTS From 282 generated items, 42 items were selected to be rated by clinicians and patients; 34 items were pilot tested and 16 items were included for Rasch analysis. The final SFInX, developed with the Partial Credit Model, contains 13 items and has the response categories: 'unable', 'partially able' and 'able'. It is unidimensional measuring 'shoulder function', and can measure from early functional use (drinking from a cup) to independence around the house (lifting items above head, carrying heavy items). CONCLUSIONS The SFInX is a promising outcome measure of shoulder function for people with a proximal humeral fracture. It has content relevant to patients and clinicians, is unidimensional and feasible for use in clinical and home settings. In its current form, the SFInX is ready for further psychometric evaluation, and for subsequent use in clinical settings and research.