Assessing risk is a key skill for all surgeons, playing a vital role in such diverse settings informed consent, audit and surgical planning. We undertook this study to define what is an acceptable level of risk for orthopaedic surgeons undertaking lower limb arthroplasty in today's medicolegal climate.

We surveyed all consultants and trainees registered with the NZOA and asked them what change in their complication rate would cause them to consider changing their practice. We also asked them how certain they would have to be that this change was real and not simply due to chance (the so-called “cluster effect”). The specific complications we considered were DVT/PE, dislocation of total hip replacement or deep infection in lower limb arthroplasty patients. This allowed us to calculate with 95% accuracy the levels of complication surgeons would deem unacceptable.

Using Cochrane's criteria for statistically valid survey results we obtained a greater than 95% representative sample. From these responses we calculated levels of complication which would worry the “average” surgeon (median value), the “vast majority” (95%) of surgeons and a change in complication rate which would be required to satisfy “statistical significance” (i.e. the level set literature proof with p<0.05 and power of 0.8). When considering deep infection following hip or knee arthroplasty: using a baseline of 0.9% as an average published rate, the average surgeon would consider changing their practice if their rate reached 2.5% and 95% of surgeons would consider changing if their rate reached 5%. For hip dislocation the baseline was set at 2.5% with the average surgeon intervening if their rate rose to 4% and the vast majority of surgeons intervening at 10%. For fatal PE the baseline was given as 0.2% with the average surgeon auditing their practice at 0.8% and the vast majority of surgeons concerned with a rate of 2.2% or higher. No difference was found in consultant/trainee responses except in the reason given for use of thromboprophylaxis where consultants used it for prevention of fatal PE and registrars for non-fatal PE.

These results could be used both as a potential guide for trainees and consultants in audit and also in guiding the use of, for instance, thromboprophylaxis where the “number needed to treat” for low molecular weight heparin to prevent one fatal pulmonary embolus in a year would necessitate national use/guidelines.