Article

Patient-reported outcome measures (PROMs) are used by some
arthroplasty registries to evaluate results after surgery,
but non-response may
bias the results. The aim was to identify a potential bias in the outcome scores
of subgroups in a cohort of patients from the Danish Shoulder Arthroplasty
Registry (DSR) and to characterize non-responders. Methods Patient-reported
outcome of 787 patients operated in 2008 was assessed 12 months postoperatively
using the Western Ontario Osteoarthritis of the Shoulder (WOOS) index. In January
2012, non-responders and incomplete responders were sent a postal reminder.
Non-responders to the postal reminder were contacted by telephone. Total WOOS
score and WOOS subscales were compared for initial responders (n = 509),
responders to the postal reminder (n = 156), and responders after telephone
contact (n = 27). The predefined variables age, sex, diagnosis, geographical
region, and reoperation rate were compared for responding and non-responding
cohorts. Results A postal reminder increased the response rate from 65% (6%
incomplete) to 80% (3% incomplete) and telephone contact resulted in a further
increase to 82% (2% incomplete). We did not find any statistically significant
differences in total WOOS score or in any of the WOOS subscales between
responders to the original questionnaire, responders to the postal reminder, and
responders after telephone contact. However, a trend of worse outcome for
non-responders was found.
The response rate was lower in younger patients.
Interpretation Non-responders did not appear to bias the overall results after
shoulder replacement despite a trend of worse outcome for a subgroup of
non-responders. As response rates rose markedly by the use of postal reminders,
we recommend the use of reminders in arthroplasty registries using PROMs.

Langue : ANGLAIS