Impact of active patient follow-up on worst-case implant survival analysis

Background: Patients lost to follow-up pose problems when performing implant survival analysis. We are unaware of any report detailing how implant survival is affected by the sequential capture of patients. Aims: To assess the logistics of long-term follow-up of patients undergoing Birmingham Hip Resurfacing (BHR), and to determine what impact this had on implant survival with time. Methods: Between 1997 and 2006 data was prospectively collected on consecutive BHR's implanted by a single surgeon in patients under 50 years of age. Survival analysis was performed at monthly intervals during this study with all patients lost to follow-up at any point in time considered to have failed. Results: There were 447 BHR's implanted in 393 patients. At study commencement cumulative 10-year survival was 42% (95% CI 36% to 48%) calculated on the confirmed knowledge of the outcome at that time of 275 BHR's, with the remaining 172 lost to follow-up and considered to have been revised. As patients were actively followed-up there was gradual improvement in survival and narrowing of confidence intervals with a final 10-year BHR survival of 96.3% (95% CI 93.7% to 98.3%) with no loss to follow-up. Conclusions: The findings demonstrate improvement in long-term implant survival with the sequential capture of young and active patients undergoing BHR who were initially considered lost to follow-up. The results reiterate the recommendation that every effort should be made to follow all patients up after joint replacement. This will improve the accuracy of survival reporting and assist comparison between different prostheses.