Background Societal and financial impacts of osteoarthritis of the hip are considerable. Weight loss, pain relief and exercise help, but total hip replacement (THR) or resurfacing (RS) are frequently necessary. In the UK 80,000 of these operations are undertaken annually costing £65m. In an ageing population, costs will escalate especially as some fail and need revision. Manufacturers market more than 100 different devices and some have recently attracted controversy. We used a new method to compare clinical and cost effectiveness of different types of THR and RS for NICE.

Methods We undertook a systematic review of randomised controlled trials (RCTs) of effectiveness supplemented with data from the National Joint Registry (NJR) for >240,000 patients, from the patient reported outcomes measures (PROMs) database and from the NHS Supply chain. We derived five frequently-used categories of THR, differing by bearing surface (ceramic, metal, polyethylene) and fixation (cemented/cementless). Probabilistic Markov modelling was used to compare lifetime cost effectiveness of THR and RS for men and women of various ages. Model outputs were mean costs and mean quality adjusted life years (QALYs) gained.

Results The systematic review yielded 37 studies indicating that outcomes differed little between interventions. RCTs were small, short and of restricted relevance. Accurate estimates of quality of life and costs were difficult to obtain. Almost all revision rates were lower than 5% at ten years (the current national benchmark is 10%). Revision rates were higher for RS than for THR, and for THR were higher for men than women. Cemented THRs with polyethylene/ceramic bearing surfaces had the lowest revision rates. Life time QALYs varied between 6.8 and 12.6 and costs between £10,781 and £28,677 according to age, gender and intervention. Overall, RS delivered fewer QALYs at higher cost. Cemented prostheses with polyethylene/metal or polyethylene/ceramic bearing surfaces were most cost effective depending on age and gender.

Conclusion RCT evidence in this field is unhelpful. Based on observational registry data, RS is not cost effective relative to THR. Cemented THR devices are more cost effective than uncemented ones. Surprisingly for such common, important operations in the NHS, accurate costs and quality of life data are not readily available. Confidential purchasing arrangements by manufacturers and hospitals hamper research. Differences in the cost effectiveness of frequently used THR types are small but important for the NHS for a large active, ageing population. The national revision rate benchmark for devices should be reduced to 5% at ten years.