Design: retrospective matched-cohort study with propensity-score matching, prognostic level III
Population: 2,358 patients aged 65 and over with operatively managed native proximal femoral fracture; those delayed for medical reasons excluded
Primary outcome: mortality at 30 and 365 days
Effect: 365-day mortality hazard ratio 1.37 (95% CI 1.15 to 1.63); 30-day 1.71 (95% CI 1.11 to 2.65)

Every audit of hip fracture care measures time to surgery, and every surgeon knows the figure is confounded: some patients wait because they need optimising, and those patients were always going to do worse. This retrospective cohort from a UK teaching hospital removed that confounder by excluding anyone delayed for medical reasons, leaving only delay caused by lack of operating capacity.

Of 3,094 screened patients, 2,358 aged 65 or over with an operatively managed native proximal femoral fracture were eligible. Capacity delay was defined as anaesthesia starting more than 36 hours after presentation to the emergency department, and it affected 1,271 (54%). After propensity-score matching — median age 85, 73% female — capacity delay carried a higher hazard of death at 365 days, hazard ratio 1.37 (95% CI 1.15 to 1.63), most pronounced in the first 30 days, hazard ratio 1.71 (95% CI 1.11 to 2.65). The effect was concentrated in frailer patients: 30-day mortality odds ratio 1.76 (95% CI 1.06 to 3.02) in those with a Clinical Frailty Scale score above 4, with no significant effect in the less frail.

That last result is the operationally useful one. It says the trauma list should be ordered by frailty as well as by waiting time, because the fit 70-year-old tolerates a day's wait in a way the frail 88-year-old does not. Most lists are ordered by arrival or by perceived technical urgency instead.

In Indian practice the constraint is usually more severe than in the cohort studied — a single trauma theatre serving a whole hospital, and a 36-hour target that is aspirational. The finding does not make capacity appear, but it does change who should go first when it is short, and it gives a number to put in front of an administrator.

Record a Clinical Frailty Scale score at admission for every hip fracture and use it to order the trauma list
Separate capacity delay from medical delay in your own audit data — pooled 'time to surgery' hides the actionable half
Prioritise the frail patient over the fit one when the list will not accommodate both
Do not let a medical optimisation label be applied to what is actually a theatre queue; it removes the delay from scrutiny
Take the 36-hour mark as the threshold this study used, not as evidence that 35 hours is safe

Why it matters

It converts a familiar audit metric into a triage rule: the queue harms frail patients specifically, and the list can be reordered today.

The statistics, in plain English

A hazard ratio of 1.37 with limits of 1.15 to 1.63 for one-year mortality is a modest but firm effect in a population with high baseline mortality, which makes the absolute difference clinically substantial. The 30-day figure is larger and less precise, 1.11 to 2.65, which is what happens when the same effect is measured over fewer events. Propensity-score matching balances the covariates that were recorded; it cannot balance the reasons a particular patient ended up in the queue, so residual confounding remains possible. The matched cohort was 98% White in a single UK hospital, so the mortality rates will not transfer directly.