Nearly 100,000 people underwent total hip replacement (THR) in the UK in 2018 and most can expect it to last at least 25 years. However, some THRs fail and require revision surgery, which results in worse outcomes for the patient and is costly to the health service. Variation in the survival of THR implants has been observed between units and reducing this unwarranted variation is one focus of the ‘Getting it Right First Time’ (GIRFT) programme in the UK. We aimed to investigate whether the statistically improved implant survival of THRs in a high performing unit is associated with the implants used or other factors at that unit, such as surgical skill.
Methods and findings
We analysed national, mandatory, prospective, cohort study (National Joint Registry for England, Wales, Northern Ireland and the Isle of Man (NJR)) of all THRs performed in England and Wales. We included the 664,761 patients with records in the NJR who have received a stemmed primary THR between 1st April 2003 and 31st December 2017 in one of 461 hospitals, with osteoarthritis as the only indication. The exposure was the unit (hospital) in which the THR was implanted. We compared survival of THRs implanted in the “exemplar” unit to THRs implanted anywhere else in the registry. The outcome was revision surgery of any part of the THR construct for any reason. Net failure was calculated using Kaplan-Meier estimates and adjusted analyses employed flexible parametric survival analysis.
The mean age of patients contributing to our analyses was 69.9 years (SD 10.1) and 61.1% were female. Crude analyses including all THRs demonstrated better implant survival at the exemplar unit with an all-cause construct failure of 1.7% (95% CI: 1.3 to 2.3) compared to 2.9% (95% CI: 2.8 to 3.0) in the rest of the country after 13.9 years(log rank test P<0.001). The same was seen in analyses adjusted for age, sex, and American Society of Anaesthesiology (ASA) score (difference in Restricted Mean Survival Time 0.12 years (95% CI 0.07-0.16) (P<0.001). Adjusted analyses restricted to the same implants as the exemplar unit show no demonstrable difference in restricted mean survival time between groups after 13.9 years (P=0.34).
A limitation is that this study is observational and conclusions regarding causality cannot be inferred. Our outcome is revision surgery and whilst important, we recognise it is not the only marker of success of a THR.
ConclusionsOur results suggest that the “better than expected” implant survival results of this exemplar centre are associated with implant choice. The survival results may be replicated by adopting key treatment decisions, such as implant selection. These decisions are easier to replicate than technical skills or system factors.