Impact of non-response and of late-response by patients in a multi-centre surgical outcome audit.

Objective: To determine the impact of non-responders on the results of a multi-centre surgical outcome audit, factors associated with failure to respond and to assess the benefits of pursuing non-responders beyond two reminders.

Design: A multi-centre prospective cohort study.

Setting: English health regions of Wessex, Mersey and Northern and South-West Thames.

Subjects: Five thousand two hundred and eighty-one (5281) consecutive cases undergoing surgery for prostate disease for a 6-month period.

Main Outcome Measures: Socio-demographic and clinical factors associated with response (both early and late) and associations of these factors with outcomes (change in symptom severity, overall outcome, and occurrence of complications) three months after surgery.

Results: An overall response rate of 82.4% was achieved. Non-responders were likely to be older and in poorer physical health than responders. Although non-response would lead to the results of an audit overestimating the benefit and underestimating the adverse effects of surgery, the size of these errors is small. Pursuit of a random sample of non-responders beyond two reminders led to 45% responding, potentially boosting the overall response rate by 8% to a total of 90.4%. However, there were few significant differences between these late responders and those who responded earlier suggesting that effort and resources expended in obtaining their response would not be justified in the routine practice of audit.

Conclusions: The results show that in a large multi-centre audit of post-operative patient perceived outcome, a response rate in excess of 80% will provide sufficiently accurate data for assessing the overall outcome of the service.