In April 2009 the National Health Service (NHS) in England embarked on a national programme to collect pre- and post-operative patient reported outcome measures (PROMs) in all patients undergoing hip replacement, knee replacement, groin hernia repair, and varicose vein (VV) surgery . Participation in the programme is mandatory for all National Health Service (NHS) and independent sector providers (hospitals and treatment centres) of NHS-funded care, although individual patients can choose not to take part. Potential uses of the data include comparisons of the performance of providers, audit of the equity of care, evaluation of the cost-effectiveness of different procedures, and estimation of the amount of health gain for populations. There are plans to extend the programme to additional surgical procedures. The use of PROMs in long-term conditions is being explored. The only previous use of PROMs on a national scale is for hip surgery in Sweden .
Pilot studies for the PROMs Programme suggested that response rates to mailed post-operative questionnaires of between 75% and 90% were achievable, dependent on the operation performed [3, 4]. Whilst these are high, rates vary between health care providers from 30% to 100%  which might give rise to biased comparisons if systematic differences exist between responders and non-responders.
Although there is an extensive literature on factors influencing response in household and longitudinal surveys, [6–8] there is less evidence on surveys of hospital patients. In England, the response to the NHS inpatient survey tends to be lower amongst men, younger patients and non-white patients . However, the evidence is more mixed for surveys of specific patient populations: some studies have reported older patients are less likely to respond [10, 11]; some have found the opposite  and others report either no association with age  or those at both extremes (youngest and oldest) are less likely to respond . As regards the sex of patients, several studies have found no association [11–14]. Most studies have found that patients in worse health at the time of their hospital admission are less likely to respond to later questionnaires [10, 11, 15, 16] although one study did not find such an association . The earlier small study we conducted of PROMs in surgery found that response was higher among older and less deprived patients .
The importance of responder bias is evident in the many studies that have reported that non-responders' have poorer outcomes [10, 11, 13–20]. Non-response therefore carries the risk of over-estimating the outcomes of providers with lower response rates. The aim of this study was to establish whether non-response is associated with patients' socio-demographic and clinical characteristics or with organisational factors for four surgical procedures. It was observational rather than testing a hypothesis.