Who gets a dual-mobility hip depends mostly on which surgeon they meet

Design

retrospective cohort study using cross-classified multilevel logistic regression on national registry data

Population

238,455 primary total hip arthroplasties in England, 2018 to 2022, of which 7,032 used dual-mobility components

Primary outcome

proportion of variation in dual-mobility use attributable to patient, surgeon and institution

Effect

surgeon 32.4%, institution 22.9%, patient factors about 10% (marginal R²=0.10, conditional R²=0.598)

Two hundred and thirty-eight thousand four hundred and fifty-five primary total hip arthroplasties from the National Joint Registry for England, linked to Hospital Episode Statistics for 2018 to 2022, were analysed with cross-classified multilevel logistic regression to work out what actually determines whether a patient receives a dual-mobility component.

Seven thousand and thirty-two received one. They were older, frailer and more often operated on for a femoral neck fracture — the selection you would expect. But the variance partition tells a different story: surgeon-level factors explained 32.4% of the variation and institutional factors 22.9%, against roughly 10% for all patient characteristics combined (marginal R²=0.10, conditional R²=0.598). Restricting the analysis to high-volume surgeons did not change the pattern.

So the strongest predictor of receiving a dual-mobility hip is not the patient's dislocation risk but the operating list they land on. Some of that is legitimate — a surgeon's judgement is not a bias — but a threefold larger contribution from surgeon than from patient is hard to read as selection on risk.

The honest conclusion is that there is no agreed indication. Without a validated dislocation-risk tool there is nothing for a surgeon to select on except experience and habit, and the authors identify exactly that gap as the prerequisite for a trial worth running. Until then, the useful action is local: look at your own unit's rate against its case mix and ask which of the two explains it.

• Audit your unit's dual-mobility rate against its case mix rather than against a national average
• State the indication in the operation note — frailty, neuromuscular disease, prior dislocation, fracture — so the decision is reviewable
• Do not read variation between colleagues as evidence that one of them is wrong; read it as an absent indication
• Follow work on dislocation-risk prediction; it is the missing piece, not implant data
• Keep counting revisions for instability locally — that is the outcome the choice is meant to affect

The statistics, in plain English

A marginal R² of 0.10 against a conditional R² of 0.598 is the whole finding in two numbers: patient characteristics explain a tenth of the variation, and adding surgeon and hospital identity raises explanatory power to nearly six-tenths. Variance partitioning tells you where variation sits, not whether it is justified — a surgeon effect could encode real expertise that the recorded patient variables do not capture. With 238,455 procedures, precision is not the limitation here; interpretation is.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.