Lost to follow-up in a hip prosthesis register. Experience of R.I.P.O.

Abstract

Patients undergoing hip arthroplasty are frequently lost to follow-up due to the generally good outcome of surgery. The lost to follow-up rate in local registers ranges from 5% (Berry et al. 1997) to 15% (Fender et al 2000). In survival analysis it is commonly assumed that ‘lost to follow-up patients’ have the same outcome as those regularly checked. This hypothesis Ž ts well with National Registers (Havelin 1999, Herberts and Malchau 2000, Lucht 2000), where the percentage of lost to follow-up is close to zero, but can be debated when applied to local registers. The hypothesis was for the Ž rst time proposed by Dorey (Dorey and Amstutz 1989), but other authors did not agree, as they proved that patients lost to follow-up had an inferior outcome than those that continue to be assessed (Murray et al. 1997), also in different orthopaedic scenarios such as shoulder surgery. (Wildner 1995). At Rizzoli Orthopaedic Institute there is a register of all total hip arthroplasty (THA) operations that goes back to 1990 (Stea et al. 2001). The register was set up in 1997, and data was collected retrospectively from 1990. Since then the register has been kept up to date. The data collected for each operation include personal information, essential clinical data, antibiotic and antithromboembolic prophylaxis, main data on surgery and complications, stick-label-based description of implant components and Ž xation. Both primary and revision operations are registered. In 11 years’ practice, the percentage of lost to follow-up has been very high (51.1% patients or 50.3% hips). This is mainly due to two reasons: Lost to follow-up in a hip prosthesis register